U.S. patent application number 14/609367 was filed with the patent office on 2015-06-25 for distributed multi-channel cognitive mac protocol.
The applicant listed for this patent is Core Wireless Licensing S.a.r.l.. Invention is credited to Kaveh GHABOOSI, Matti LATVA-AHO.
Application Number | 20150181562 14/609367 |
Document ID | / |
Family ID | 41161582 |
Filed Date | 2015-06-25 |
United States Patent
Application |
20150181562 |
Kind Code |
A1 |
GHABOOSI; Kaveh ; et
al. |
June 25, 2015 |
DISTRIBUTED MULTI-CHANNEL COGNITIVE MAC PROTOCOL
Abstract
A method includes sending a message from a first to at least one
second cognitive radio apparatus to determine a channel to be used
for sending data from the first to the second radio apparatus, the
message sent over a first channel and comprising an advertisement
of at least one second channel for use in sending the data from the
first to the second radio apparatus, the advertisement comprising a
corresponding proposition/evaluation bit for each second channel,
receiving a reply from the second radio apparatus over the first
channel, comprising an acceptance of one of the second channels
with the corresponding proposition/evaluation bit, a rejection of
the second channel and an advertisement of a third channel, or a
rejection of the second channel without an advertisement of a third
channel, and transmitting the data from the first to the second
radio apparatus over an agreed upon second or third channel.
Inventors: |
GHABOOSI; Kaveh; (Oulu,
FI) ; LATVA-AHO; Matti; (Oulu, FI) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Core Wireless Licensing S.a.r.l. |
Luxembourg |
|
LU |
|
|
Family ID: |
41161582 |
Appl. No.: |
14/609367 |
Filed: |
January 29, 2015 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
14450281 |
Aug 3, 2014 |
8971943 |
|
|
14609367 |
|
|
|
|
12082361 |
Apr 9, 2008 |
8831519 |
|
|
14450281 |
|
|
|
|
Current U.S.
Class: |
370/329 |
Current CPC
Class: |
H04W 16/14 20130101;
H04W 84/12 20130101; H04W 74/002 20130101; H04W 74/08 20130101;
Y02D 70/142 20180101; H04W 72/04 20130101; Y02D 30/70 20200801;
H04W 72/0406 20130101; Y02D 70/22 20180101; H04W 52/0206
20130101 |
International
Class: |
H04W 72/04 20060101
H04W072/04 |
Claims
1. A method, comprising; sending a message from a first cognitive
radio apparatus to at least one second cognitive radio apparatus
during a negotiation phase to determine a channel to be used for
sending data from the first cognitive radio apparatus to the at
least one second cognitive radio apparatus, the message being sent
over a first communication channel and comprising an advertisement
of at least one second communication channel for use in sending the
data from the first cognitive radio apparatus to the at least one
second cognitive radio apparatus, the advertisement comprising a
corresponding proposition/evaluation bit for each of the at least
one second communication channel; receiving a reply from the at
least one second cognitive radio apparatus over the first
communication channel, the reply comprising one of an acceptance of
one of the at least one second communication channels with the
corresponding proposition/evaluation bit from the advertisement, a
rejection of the at least one second communication channel and an
advertisement of at least one third communication channel, and a
rejection of the at least one second communication channel without
an advertisement of at least one third communication channel; and
transmitting the data from the first cognitive radio apparatus to
the at least one second cognitive radio apparatus over an agreed
upon one of the second or third communication channels in response
to a completion of the negotiation phase.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application is a continuation of U.S. application Ser.
No. 14/450,281, filed 3 Aug. 2014, which is a continuation of U.S.
application Ser. No. 12/082,361, filed 4 Sep. 2008, the disclosures
of which are incorporated herein by reference in their
entirety.
FIELD
[0002] The exemplary and non-limiting embodiments disclosed herein
relate generally to wireless communication systems, methods,
devices and computer programs and, more specifically, relate to
cognitive radio apparatus and to wireless systems that are operable
with cognitive radio apparatus.
BACKGROUND
[0003] Opportunistic radio resource management (RRM) schemes have
recently received extensive attention in scientific literature and
technological fields. One possible area for which opportunistic RRM
can be effective is IEEE 802.11-based wireless local area networks
(LANs). Existing IEEE 802.11 based systems suffer from inefficient
medium access strategies, namely distributed coordination function
(DCF), point coordination function (PCF), and their corresponding
amendment supporting quality of service (QoS). Furthermore, it is
expected that problems will arise due to lack of sufficient
frequency opportunities, due at least to the fact that frequency
regulations have not efficiently allocated diverse frequency bands.
Consequently, cognitive radio, frequency agile, and opportunistic
RRM schemes aim to address the aforementioned critical problems in
an optimized fashion, resulting in better spectrum utilization and
fair radio resource allocation to associated wireless entities.
SUMMARY
[0004] In a first non-limiting aspect thereof the exemplary
embodiments provide a method that includes sending a message from a
first cognitive radio apparatus to at least one second cognitive
radio apparatus, the message being sent over a first communication
channel and comprising an advertisement of at least one second
communication channel for use in sending data from the first
cognitive radio apparatus to the at least one second cognitive
radio apparatus. The method further includes receiving a reply from
the at least one second cognitive radio apparatus over the first
communication channel, where the reply comprises one of an
acceptance of one of the at least one second communication
channels, a rejection of the at least one second communication
channel and an advertisement of at least one third communication
channel, or a rejection of the at least one second communication
channel without an advertisement of at least one third
communication channel. The method further includes transmitting the
data from the first cognitive radio apparatus to the at least one
second cognitive radio apparatus over an agreed upon one of the
second or third channels.
[0005] In another non-limiting aspect thereof the exemplary
embodiments provide a computer-readable medium that stores program
instructions, the execution of the program instructions resulting
in operations that comprise sending a message from a first
cognitive radio apparatus to at least one second cognitive radio
apparatus, the message being sent over a first communication
channel and comprising an advertisement of at least one second
communication channel for use in sending data from the first
cognitive radio apparatus to the at least one second cognitive
radio apparatus; receiving a reply from the at least one second
cognitive radio apparatus over the first communication channel, the
reply comprising one of an acceptance of one of the at least one
second communication channels, a rejection of the at least one
second communication channel and an advertisement of at least one
third communication channel, or a rejection of the at least one
second communication channel without an advertisement of at least
one third communication channel; and transmitting the data from the
first cognitive radio apparatus to the at least one second
cognitive radio apparatus over an agreed upon one of the second or
third channels.
[0006] In a further non-limiting aspect thereof the exemplary
embodiments provide a first transceiver for communication over a
first communication channel; a second frequency agile transceiver
for communication over second and third communication channels and
a controller configurable to operate the apparatus as a first
cognitive radio apparatus and to transmit a message to at least one
second cognitive radio apparatus. The message is transmitted over
the first communication channel and comprises an advertisement of
at least one second communication channel for use in sending data
from the first cognitive radio apparatus to the at least one second
cognitive radio apparatus. The controller is further configurable
to receive a reply from the at least one second cognitive radio
apparatus over the first communication channel, the reply
comprising one of an acceptance of one of the at least one second
communication channels, a rejection of the at least one second
communication channel and an advertisement of at least one third
communication channel, or a rejection of the at least one second
communication channel without an advertisement of at least one
third communication channel. The control unit is further
configurable to transmit the data to the at least one second
cognitive radio apparatus over an agreed upon one of the second or
third channels.
BRIEF DESCRIPTION OF THE DRAWINGS
[0007] In the attached Drawing Figures:
[0008] FIG. 1A shows a high level block diagram of a cognitive
radio system having a plurality of STAs and a BSS;
[0009] FIG. 1B shows a simplified block diagram of a
two-transceiver STA;
[0010] FIG. 2A shows the current channel numbering scheme of IEEE
802;11 a, FIG. 2B shows the case of the IEEE 802;11b and IEEE
802;11g, 2;4 GHz band, while FIG. 2C shows the IEEE 802;11y 3 GHz
band;
[0011] FIG. 3 shows a position of a new Channel Information (CI)
field in control frames;
[0012] FIG. 4 shows a generic format of the Channel Information
(CI) field;
[0013] FIG. 5 shows a generic sub-field format of the Channel
Control (CC) field shown in FIG. 4;
[0014] FIGS. 6, 6A and 6B show Reason Code Bit Patterns that
pertain to the Reason Code sub-field of FIG. 5;
[0015] FIG. 7 illustrates possible frame exchange scenarios for PS
mode enabled and PS mode disabled cases;
[0016] FIG. 8 illustrates Channel Control (CC) inclusion within the
exchanged control frames;
[0017] FIG. 9 shows a simple example of a task list;
[0018] FIGS. 10A-10D describe four modes I, II, Ill and IV,
respectively, that are related to a non-ISM band convergence
concept;
[0019] FIG. 11 shows a "Reception" task merging procedure for the
Task List shown in FIG. 1B;
[0020] FIG. 12 shows a "Transmission" task merging procedure for
the Task List shown in FIG. 1B;
[0021] FIG. 13 illustrates sub-domains of a LSEChannel Time Index
Domain;
[0022] FIG. 14 shows an Information Element for Primary User
Appearance Reporting;
[0023] FIG. 15 illustrates a Channel Information Database,
Management Entities, and Transient Zone;
[0024] FIG. 16 shows a Task List maintenance rule and strategy that
accommodates a single transceiver STA embodiment;
[0025] FIG. 17 shows a simplified block diagram of a
two-transceiver cognitive mesh entity (CME);
[0026] FIG. 18 is a depiction of a Common Channel Framework (CCF)
concept in IEEE 802;11s;
[0027] FIG. 19 shows an example of a permanent channel switching
announcement using a CHSW frame transmitted on the shared ISM
channel;
[0028] FIG. 20A shows the eRTX frame format in a cognitive common
channel framework (CCCF), while FIG. 20B shows the eCTX frame
format;
[0029] FIG. 21 shows a Channel Switching (CHSW) frame format in the
CCCF;
[0030] FIG. 22 depicts a channel switching information element
(CHSWIE) structure in eRTX, eCTX, and CHSW frames;
[0031] FIGS. 23A and 23B are tables showing a Reason Code Bit
Pattern (CHSWIE Status=11) and a Reason Code Bit Pattern (CHSWIE
Status=00);
[0032] FIG. 24 shows a channel reservation using eRTX and eCTX, for
a case where the knowledge of the CSE concerning the LTRC of the
CDE is correct;
[0033] FIG. 25 shows a channel reservation using eRTX and eCTX for
a case where only the LTRC channel information of the CDE is
included in the eRTX, while the knowledge of the CSE of the LTRC of
the CDE is incorrect;
[0034] FIG. 26 shows the channel reservation using eRTX and eCTX
when the CSE has no a priori knowledge of the LTRC of the CDE;
[0035] FIG. 27 shows the channel reservation using eRTX and eCTX
when the knowledge of the CDE of the LTRC of the CSE is
incorrect;
[0036] FIG. 28 illustrates the eRTX CHSWIE possible configurations
and the corresponding Duration/ID and OLD sub-field setup;
[0037] FIG. 29 illustrates the eCTX CHSWIE possible configurations
and the corresponding Duration/ID and OLD sub-field setup;
[0038] FIG. 30 illustrates in detail the frame format of a SWinv
control frame and its associated CHSWIE;
[0039] FIG. 31 illustrates exemplary message flow for a multicast
temporary channel switching use case;
[0040] FIG. 32 illustrates exemplary message flow for a Mode I
multicast permanent channel switching use case;
[0041] FIG. 33 illustrates exemplary message flow for a Mode II
multicast permanent channel (fast) switching use case;
[0042] FIG. 34 illustrates exemplary message flow for a combined
multicast/unicast channel switching use case; and
[0043] FIG. 35 is a logic flow diagram that illustrates the
operation of a method, and a result of execution of computer
program instructions, in accordance with the exemplary
embodiments.
DETAILED DESCRIPTION
[0044] The exemplary embodiments provide a novel distributed
frequency agile medium access control (MAC) protocol suitable for
use in next generation wireless LANs, which furthermore have
complete backwards compatibility with the legacy 802.11 systems.
The enhanced MAC protocol is capable of multi-channel deployment of
available frequency bands to coordinate concurrent multiple data
transmissions. Previous multi-channel MAC protocols for wireless
LANs that have been proposed in the literature generally make
unreasonable assumptions, while being unable to address technical
problems concerning channel utilization and simultaneous
information transmissions. In contrast, the exemplary embodiments
provide an optimized MAC protocol, which is capable of addressing a
variety of problems inherent in multi-channel systems, while being
aware of primary users in different frequency bands using
intelligent environmental information management entities.
[0045] The use of the exemplary embodiments also improve the
channel utilization and capacity using the concept of cognitive
radio and also reduce access delay due to more intelligent decision
making procedures used for link layer connection establishment. In
addition, a concept of welfare enhancement (WE) is provided, that
results in higher channel utilization and system throughput. Future
extensions to the disclosed enhanced protocols may be simply
incorporated, without requiring major protocol core code
modification.
[0046] New fields for control/management frames are provided in
such a way that legacy STAs are able to understand legacy fields
while discarding the newly added fields (and sub-fields). Legacy
STAs are able to receive and decode all frames generated by
cognitive "smart" STAs except for those fields dedicated and
particularly designed for cognitive STAs. Thus, both legacy and new
systems are able to work with each other. Legacy STAs are able to
join cognitive BSSs and vice versa.
[0047] FIG. 1A shows a high level block diagram of a cognitive
radio system having a plurality of stations (STAs) 10, a basis
service set (BSS) 12 and a cognitive basis service set (CBSS) 14.
Certain of the STAs 10 may be cognitive STAs 10A, and others may be
legacy (non-cognitive) STAs 10B.
[0048] FIG. 1B shows a simplified block diagram of a cognitive STA
10A. In this example the cognitive STA 10A includes two wireless
transceivers, specifically an ISM band (industrial, scientific,
medical band) transceiver 20 and a cognitive transceiver 22. The
STA 10A also includes a controller, such as at least one data
processor (DP) 21 that operates in accordance with a stored program
in a memory 23. Execution of the program instructions results in
the implementation of a MAC entity 26 that is constructed and
operated in accordance with the exemplary embodiments. The
cognitive STA 10A also includes transceiver associated task lists
24A, 24B (which may be collectively referred to as a task list 24),
and other components and data structures as described in detail
below.
[0049] Legacy IEEE 802.11 based wireless networks are currently
capable of operation in certain dedicated ISM channels. As such, a
need exists to enable next generation wireless LANs to operate in
any frequency band, while being able to work and interact with
existing legacy 802.11 networks. For example, assume a case of an
established legacy 802.11 basic service set (BSS) 12 comprising
associated 802.11 stations (STAs) 10 working/collaborating with
another cognitive basic service set (CBSS) 14 which is able to
provide additional network connectivity and packet forwarding to
the legacy system. Further in this regard providing additional
network connectivity and packet forwarding to the legacy BSS 12
mandates that the CBSS 14 operate at the same frequency band and
ISM channel as the existing legacy 802.11 system. This fact implies
that legacy BSS 12 and CBSS 14 may be better combined with one
another to produce a global widespread 802.11 BSS. As a result, and
in the same ISM channel, there is provided a BSS 12 to which both
legacy and cognitive STAs 10 are associated and cooperating with
each other in different ways, e.g., packet forwarding, frame
buffering, etc.
[0050] On the other hand, it is desirable to not congest the shared
ISM channel with those frames exchanged by cognitive STAs 10A. This
becomes possible when cognitive STAs 10A utilize non-ISM data
channels other than the shared channel, which is utilized
particularly by legacy 802.11 STAs 10. As a result, one may
conclude that it is better for the shared ISM channel to be
utilized predominantly by legacy STAs 10B for management, control,
and data exchange purposes, while the shared ISM channel is used
only for management and control purposes for cognitive STAs
10A.
[0051] According to the foregoing, since both legacy and cognitive
STAs 10 can be associated to a common BSS 12 while using the same
ISM channel for management/control purposes, consequently they are
able to cooperate in control and management functions. Thus, it can
be argued that from management/control point of view, there is no
difference between a legacy 802.11 STA 10B and a cognitive 802.11
STA 10A. On the other hand, and as pointed out, cognitive STAs 10A
should preferably not use the shared ISM channel for their data
communication. As a result, cognitive STAs 10A are strongly
mandated to switch to another channel (referred to herein as a
non-ISM channel) in order to perform their information exchange.
However, a problem that arises is how the receiving STA 10 knows to
which channel it should switch to enable the source STA 10 to
transmit MSDU(s). In addition, a related problem is how the source
STA 10 recognizes in which channel its intended receiver is waiting
to receive its targeted MSDU(s).
[0052] The cognitive destination STA 10A should be able to
determine and switch to the intended channel, if prior to channel
switching both source and destination STAs 10 agreed to a channel
to be used during their data communication. In other words, the
source and destination STAs 10 should preferably negotiate
concerning the channel to be utilized for the upcoming
communication beforehand and reach a mutual agreement. Hence, the
shared ISM channel is used to exchange control information in the
form of RTS, CTS, ATIM, ACK, etc., to let both involved parties
advertise their desired channel(s). Upon reaching mutual agreement,
the parties (STAs 10) switch to the agreed channel simultaneously
and start their data communication. As a result, by the use of this
approach there is no need for a separate dedicated control channel
for management/control purposes.
[0053] It is thus preferred that both legacy and cognitive STAs 10A
share a common ISM channel for both management and control
purposes. Technically, there is no difference between BSSs 12
established by legacy STAs 10B and the BSSs 12 initially created by
cognitive STAs 10A. In addition, legacy STAs 10B should be able to
join BSSs 12 formerly established by cognitive STAs 10A, and vice
versa. In any situation, and whether the BSS 12 was initiated
either by a cognitive STA 10A or a legacy STA 10B, cognitive radios
are the guests to the ISM band. As a result, they provide
additional services to the legacy STAs 10B which are licensed users
in the band. The basic services offered by cognitive STAs 10A may
include, but are not limited to, network connectivity and packet
forwarding in layer three. On the other hand, it is preferred that
cognitive STAs 10A are only allowed to use common ISM channel to
establish their connection and reach agreement concerning the
channel to be utilized for their data communication.
[0054] Information regarding primary user appearance and related
issues are handled using an extended version of the dynamic
frequency selection (DFS) scheme proposed in IEEE 802.11 h. By
combining the existing DFS scheme with a number of new concepts, a
novel extension to DFS is provided that is capable of addressing
problems related to primary user appearance in non-ISM
channels.
[0055] The exemplary embodiments also provide novel approaches to
preventing undesired congestion in non-ISM channels. Also, by
proposing another class of concepts, channel aggregation/bursting
in multi-channel scenarios becomes possible, resulting in an
enhancement in channel utilization.
[0056] In the ensuing description channel identification approaches
are described that are suitable for use in broadband cognitive
wireless local area networks, followed by a description of an
enhanced medium access control protocol for cognitive wireless
LANs.
Channel Identification Approaches
[0057] When two cognitive STAs 10A desire to communicate with one
another they should reach a reciprocal conformation concerning the
non-ISM channel to be utilized during their data communication. As
explained above the common ISM channel, in which the BSS 12 has
been formed, is particularly used for connection establishment and
control frame exchange. During this phase, both source and
destination STAs 10 should negotiate regarding their desired
non-ISM channels to be occupied for the whole data exchange.
However, the way in which STAs 10 advertise their intended channels
is preferably dealt with explicitly. While cognitive STAs 10A may
be mandated to use common approaches when advertising desired
channels, this can be done in several ways.
[0058] One simple way to advertise intended channels is to mention
their channel identifiers (CIs) within exchanged control frames
(e.g., RTS, CTS, etc.). However, there is no globally unique and
wholly accepted channel-numbering scheme. For example, in IEEE
802.11a the current channel numbering scheme (Channel Identifiers)
are as shown in FIG. 2A, while FIG. 2B shows the channel numbering
case of IEEE 802.11b and IEEE 802.11g, 2.4 GHz band, while FIG. 2C
shows the channel numbering scheme of the IEEE 802.11y 3 GHz
band.
[0059] As should be evident, there is no global agreement as to how
different center frequencies are numbered and, as a result,
inclusion of the center frequency of an advertised channel is
preferred as well, and the inclusion of a "channel identifier" for
preferred non-ISM channel(s) is not necessary.
[0060] In IEEE 802.11y, in addition to channel center frequency,
the corresponding regulatory class is also defined. Basically, the
regulatory class determines several physical layer transmission
specifications, including channel starting frequency, channel
spacing, channel set, transmission power limit, emissions limits
set, etc. As a result, in addition to center frequency, it would be
desirable to include the regulatory class when advertising intended
channel(s). In addition, the inclusion of the regulatory class when
advertising intended channel(s) is preferable to only mentioning
the targeted channel bandwidth. Further in this regard, in the
future not only conventional channel bandwidths, e.g., 5, 10, and
20 MHz, will be available, but also other channel bandwidths may be
allowed by regulatory authorities. Therefore, a channel bandwidth
field may also be included within control frames during the channel
negotiation phase. As a result, in addition to the channel center
frequency the inclusion of both a regulatory class field and a
channel bandwidth field are desirable.
Primary User Detection
[0061] There is no major problem presented by Primary User
Appearance in the common ISM channel shared by legacy and cognitive
STAs 10A. This due at least to the fact that IEEE 802.11-based
systems are licensed in these frequency bands, and as a result,
spectrum sensing for primary user detection in ISM bands is not
mandatory, except for the case of IEEE 802.11a (which should be
aware of radar signals). Therefore, in a BSS 12 comprising both
legacy and cognitive STAs 10A, if the shared ISM channel is in the
5 GHz band, all associated STAs 10, including legacy and cognitive
STAs 10A, should cooperate in the dynamic frequency selection (DFS)
procedure. On the other hand, for all discovered non-ISM frequency
opportunities that may be utilized by cognitive STAs 10A it is
preferred that there be regular (and possibly randomized) spectrum
sensing activities to assure very limited interference to primary
users of this spectrum. Those non-ISM channels in which the primary
user is detected are preferably recorded accordingly by all
involving STAs 10. In addition, it is desirable that there be a set
of counters for every channel in which primary user has been
detected to prevent the MAC entity 26 from acquiring forbidden
channels. The manner by which cognitive STAs 10A cooperate in
spectrum sensing may be based on an extended version of the IEEE
802.11h DFS scheme. The considered modulation scheme to be
implemented in the physical layer is based on orthogonal frequency
division multiplexing (OFDM). As a result, any non-decodable
signaling in non-ISM channels may be considered as a Primary User
Appearance indication. Basically, if clear channel assessment
(CCA), which is a combination of carrier sensing (CS) and energy
detection (ED), detects any type of known Clause 19 synchronization
(sync) symbols, including Barker code sync and OFDM sync symbols,
it can be concluded that the detected carrier is due to a
legacy/cognitive IEEE 802.11 STA 10; otherwise, it may be concluded
that the detected carrier is due to a primary user and subsequently
the occupied non-ISM channel is released. In general, the cognitive
STAs 10A perform spectrum sensing in at least two different ways: a
randomized approach and a regular approach.
Frame Exchange Structure for the Cognitive MAC
[0062] In order to advertise a certain channel to be used during
data communication between a pair of STAs 10 the MAC entity 26,
depending on the current situation, may send an ATIM/RTS/Negative
non-Null CTS/Negative non-Null ACK frame to its intended STA MAC
entity 26.
[0063] More specifically, the Announcement Traffic Indication
Message (ATIM) is used to inform the recipient about any pending
MSDU addressed to it during a so-called ATIM window just after
Beacon frame transmission. Upon reception of an ATIM, the recipient
responds with an ACK frame if it agrees to receive the buffered
MSDU(s). On the other hand, if the intended recipient has left the
IBSS, there will be no respond and as a result, the buffered
MSDU(s), after expiration of buffering timeout, are discarded.
Basically, ATIM/ACK is used when a power saving (PS) mode is
enabled in an IBSS (or BSS). Here, in the cognitive MAC protocol of
interest to these exemplary embodiments, an ATIM frame may have the
following frame formats as its subsequent response:
[0064] regular ACK (when the recipient is a legacy STA, it replies
the received ATIM frame using a conventional ACK frame, as in IEEE
802.11),
[0065] cognitive Positive ACK (when the recipient is a cognitive
STA 10A, and if it agrees with at least one of the
offered/advertised channels, it responds with the Positive
ACK),
[0066] cognitive Negative Null ACK (when the recipient is a
cognitive STA, and if the recipient disagrees with all
offered/advertised channels and has no suggestion to offer, it
responds with the Negative Null ACK), and
[0067] cognitive Negative non-Null ACK (when the recipient is a
cognitive STA, and if it disagrees with all offered/advertised
channels and has a better suggestion(s) to offer, it responds with
the Negative Null ACK that includes its suggestion(s)).
[0068] In the latter case the source STA 10 may respond using the
Positive or the Negative Null ACK.
[0069] The Ready-To-Send (RTS) is used to inform the recipient
about any pending MSDU addressed to it. In the cognitive MAC entity
26 protocol in accordance with these exemplary embodiments an RTS
frame can have the following frame formats as its subsequent
response:
[0070] regular CTS (when the recipient is a legacy STA 10B, it
replies to the received RTS frame using a conventional ACK frame,
as in IEEE 802.11),
[0071] cognitive Positive ACK (when the recipient is a cognitive
STA 10A, and if it agrees with at least one of the
offered/advertised channels, it responds with a Positive CTS),
[0072] cognitive Negative Null CTS (when the recipient is a
cognitive STA 10A, and if it disagrees with all offered/advertised
channels while having no channel suggestion of its own to offer, it
responds with a Negative Null CTS), or
[0073] cognitive Negative non-Null CTS (when the recipient is a
cognitive STA 10A, and if it disagrees with all offered/advertised
channels while having a (better) channel suggestion(s) to offer, it
responds with a Negative non-Null CTS, and places its suggestion(s)
within the Negative non-Null CTS). In this case, the source STA may
respond using the Positive or Negative Null ACK).
[0074] After performing the regular DCF access procedure (for the
case of ATIM and RTS) the source STA 10 may transmit an ATIM/RTS
frame to the intended destination STA. The size of the transmitted
ATIM/RTS frame is preferably considered to be variable since it is
not clear how many channels are going to be advertised to the
intended receiver. The transmitted ATIM/RTS has the following
fields: Frame Control (2 Bytes), Duration/ID (2 Bytes), Receiver
Address (6 Bytes), Transmitter Address (6 Bytes), BSSID (Basic
Service Set ID) (6 Bytes), Sequence Control (2 Bytes) (Only for
ATIM), Frame Check Sequence (FCS) (4 Bytes), and finally a variable
size Channel Information (CI) field. FIG. 3 shows the position of
the Channel Information (CI) field in control frames, while FIG. 4
shows a generic Format of Channel Information (CI) Field.
[0075] In FIG. 4 the Center Frequency (CF) field (2 Bytes) is
specified in MHz, the Regulatory Class (RC) field (1 Byte) is
considered mandatory; and if not used is loaded with some default
value, such as 00 hex; and the Channel Mask field (in particular
circumstances Bandwidth (BW)) (1 Byte) which, if the RC field is
not used, is set to the recommended bandwidth; 5, 10, or 20 MHz
(e.g. 5 MHz is represented by 05H or 00000101B). Note that other
values may be used depending on the allowed channel bandwidths.
Note also that BW information may be included in the RC field, as
most regulatory classes define the bandwidth of the intended
channel, in addition to allowed EIRP, etc. Therefore, when the RC
field is used, it explicitly determines the bandwidth, EIRP (e.g.
Transmit Power Limit in mW/MHz), etc. and as a result, BW field may
not be needed. This implies that the RC is a mandatory field while
BW is non-compulsory. On the other hand, when the RC field is not
used (loaded by 00H), then BW field is used in order to determine
the bandwidth of a negotiated data channel.
[0076] The Channel Control (CC) (1 Byte) field in FIG. 4 has the
following sub-fields (reference is made to FIG. 5).
[0077] 1) A Channel Preference Number (CPN) (2 Bits) which is used
to number offered channels: 01, 10, or 11 (00 may be reserved and
is not used). Upon returning the evaluation results of an offered
channel, the CPN sub-field should have the same number as it had
when the channel was being advertised.
[0078] 2) A Proposition/Evaluation Bit (1 Bit) which is used to
determine whether the channel with a preference number included in
the CPN sub-field is a channel being advertised for the first time,
or a channel which has been previously offered and subsequently
evaluated by the other MAC entity 26, where Proposition=0 (the
channel is being advertised), Evaluation=1 (the channel evaluation
results are being disclosed).
[0079] 3) A Decision Bit (1 Bit) which is used when the
Proposition/Evaluation Bit is 1. In this case the Decision Bit
determines whether the offered channel has been accepted or
rejected: Accept=1, Reject=0 (when Proposition/Evaluation=0 this
bit is simply ignored, and by default may be set to 0).
[0080] 4) Reserved (1 Bit)
[0081] 5) A Reason Code (3 Bits) which when offering a channel
(Proposition/Evaluation Bit=0) specifies the reason for which the
channel is being advertised, as indicated in FIG. 6 (only 000, 001,
010, 011); on the other hand, when Proposition/Evaluation Bit=1
this sub-field specifies the reason for which the channel has been
accepted (only 010, 011)/rejected (only 000, 001, 100, 101, 110,
111). Note that the Decision Bit may be combined with the Reason
Code according to the aforementioned definitions.
[0082] Note that in an ATIM/RTS (traffic initiator) control frame,
there is no need to determine the duration of intended channel
deployment. Upon any conformity, both source and destination STAs
10 switch to the agreed channel and, consequently, they exchange
regular RTS/CTS to reserve the channel for a particular time
duration.
[0083] Also, the following points should be noted. First, only one
of the advertised channels might be selected at any step of the
channel negotiation phase. Second, the source (and destination) STA
10 is preferably limited in the number of advertised channels. As
an example, by default the maximum number of advertised channels
may be three (3). If the source STA 10 has more than three
candidate channels it selects the best three. On the other hand, if
the source STA 10 has less than three channels to advertise it may
present the information of only the available channels.
[0084] Referring generally to FIG. 7, when a cognitive STA 10A
wants to invite its intended destination STA to establish a
connection, it uses an ATIM or a RTS frame and puts its desired
non-ISM channel(s) within the CI field (see FIGS. 3-5). Note that
due to a desire to maintain backwards compatibility it is preferred
to use legacy RTS, CTS, ATIM, and ACK frames to enable legacy STAs
10B to receive and recognize the captured frames. If the
destination STA 10 agrees with one of the advertised channels, it
will respond with a Positive ACK or Positive CTS frame (or simply
with an ACK/CTS frame with an extra CI field as its payload). On
the other hand, when the destination STA 10 disagrees with the
offered channels, and has its own suggestions, it responds with a
Negative non-Null ACK or Negative non-Null CTS. When the receiver
(the destination STA 10 in this case) disagrees with the advertised
channels, and has no channels at hand to offer, it responds with a
Negative Null ACK or Negative Null CTS. In this case, no connection
with the intended STA 10 is established and, later, the source STA
10 may retry with another possible channel or channels. The manner
in which STAs 10 offer the channel(s) and accept/reject particular
channels is based on the following rules.
[0085] A. When a certain channel is being advertised, the promoting
STA 10 assigns a preference number to the channel and places it in
the Channel Preference Number (CPN) field (see FIG. 5). When the
recipient STA 10 evaluates the channel(s) and reports the
evaluation results, it uses the same number as the channel(s) had
when advertised.
[0086] B. When a channel is being advertised, the corresponding
Proposition/Evaluation Bit is set to 0 (Proposition), and when its
evaluation result is being reported the Proposition/Evaluation Bit
is set to 1 (Evaluation). For a candidate channel being advertised,
the Decision Bit is simply ignored (and may be set to 0). In
addition, when a channel is being offered, its physical
specifications (e.g., its regulatory class, center frequency,
bandwidth, etc.) are indicated in the Channel Information (CI)
field. The CC field of a particular channel which is being
advertised is followed immediately by its corresponding CI field.
Accordingly, the MAC entity 26 checks the Proposition/Evaluation
Bit and if it is 0 (Proposition), it expects a CI field to follow
the CC field. On the other hand, in a CC field being analyzed, when
the Proposition/Evaluation Bit is set to 1, the MAC entity 26
expects to find the evaluation report of the channel with the
preference number indicated in the CPN field in the following
Decision Bit. If the Decision Bit is 0, the MAC entity 26 can
assume that the channel was rejected by the recipient, and if it is
1, then the MAC entity 26 may conclude that the channel has been
accepted. In the case of channel evaluation reporting, the MAC
entity 26 does not expect a CI field to follow the CC field being
processed. By the use of this procedure and rule set there is no
need to include additional header(s) to express the number of
advertised/reported channels and so forth.
[0087] FIG. 8 illustrates Channel Control (CC) inclusion within the
exchanged control frames.
[0088] Upon transmission of an ATIM/RTS frame (in cognitive
format), the intended receiver analyzes the included fields to
deduce all required information. Here, at least two distinct
scenarios are possible.
[0089] First, the receiver (intended destination STA) is not a
cognitive radio. In this case, it fetches the related information
by the use of only known fields and ignores the remaining field(s)
without any further processing. The receiver responds to the
received ATIM frame with a regular ACK. Upon reception of the ACK
frame, the source STA concludes that the intended receiver is a
non-cognitive STA and, as a result, it uses the conventional ISM
band, in which current IBSS has been already established, in order
to exchange pending MSDU(s) with the receiver.
[0090] In a second scenario the receiver is a cognitive radio. In
this case there are multiple possible scenarios that can occur.
[0091] A) When the destination STA agrees with at least one of the
offered channels, it informs the source STA of its agreement to
deploy the intended channel. Note that the receiver chooses only
one of the offered channels based on its own localized decision
making procedures, which may be based upon an advanced channel
feature measurement or other performance criteria. The protocol by
which the receiver calculates the channel performance metrics can
be any suitable protocol, and is assumed to be based upon PHY and
MAC cooperation. The receiver prepares an ACK/CTS frame to convey
its agreement about one of the advertised channels. The appended CC
field within the ACK/CTS frame contains a CPN field loaded by the
preference number of the channel that was present when it was being
advertised (01, 10, or 11, where 00 may be reserved and not used),
Proposition/Evaluation Bit set to 1 (Evaluation), Decision Bit set
to 1 (Accepted), Reserved Bit, and Reason Code set to one of the
two possibilities, 010 or 011.
[0092] B) When the receiver disagrees with the offered channels,
and has no suggestion regarding the channel deployment, it sends
the Negative Null ACK/Negative Null CTS frame back to the source
STA. The appended CC field within the Negative Null ACK/Negative
Null CTS frame contains the CPN field loaded by the preference
number of the channel that was present when it was being advertised
(01, 10, or 11, where 00 may be reserved and not used),
Proposition/Evaluation Bit set to 1 (Evaluation), Decision Bit set
to 0 (Rejected), Reserved Bit, and Reason Code set to one of four
possibilities, 100, 101, 110, or 111.
[0093] C) When the receiver disagrees with the offered channels,
and has at least one other suggestion, it advertises its own
channels just as the source STA advertised its channels. In this
case then the receiver sends back a Negative non-Null ACK/Negative
non-Null CTS frame with the following fields: a set of CC fields
corresponding to the rejected channel(s), and a set of CC fields
related to the channels being advertised accompanied with their CI
fields. For the first set, the Proposition/Evaluation Bit is set to
1 while for the second set this bit is set to 0. The MAC entity 26,
by the use of this bit, is able to readily determine whether any CI
field occurs immediately following a CC field. Also, for the
rejected channels, the Reason Code may be set to one of six
possibilities, 000, 001, 100, 101, 110, or 111. Upon reception of
the Negative non-Null ACK/Negative non-Null CTS frame, at least two
different operational modes can ensue.
[0094] In a first mode, if the source STA disagrees with the
offered channels, it will send a Negative Null ACK frame with a
sufficient number of CC field(s) equivalent to the channels
advertised by the destination STA. In this case all
Proposition/Evaluation Bits in all aforementioned fields should be
1 (Evaluation), while the Decision Bits are set to 0 (Rejected).
Due to the fact that all Proposition/Evaluation Bits are 1, there
will be no CI field.(since nothing is being advertised).
[0095] In a second mode, if the source STA agrees with at least one
of the offered channels, it will inform the destination STA of its
agreement to deploy the intended channel. In this case the source
STA prepares an ACK frame to convey its agreement concerning one of
the advertised channels. The appended fields are the CC fields
corresponding to the rejected channels, with Reason Code sub-fields
set appropriately, and a single CC field that carries a CPN of the
accepted channel.
[0096] A discussion is now made of MAC entity 26 basic
functionality.
[0097] It is first noted that there are numerous operational cases
that may occur during the deployment of a cognitive wireless
network that tend to make the MAC entity 26 more complex than a
legacy/regular single channel MAC protocol (e.g., IEEE 802.11
MAC).
[0098] In general, designing a MAC algorithm for a cognitive
multi-channel system may take into account how many transceiver(s)
are available for the designer at the PHY level. In addition,
different scenarios exist in the PS and non-PS modes which should
be investigated in detail and addressed separately. One can
categorize different cases according to diverse differentiation
criteria. The number of available transceiver(s) is one of the key
classification criterions. In addition, as mentioned above being in
PS mode (PS enabled) or non-PS mode (PS disabled) is another
complicating factor. Here, and for simplicity, it is preferred to
categorize the possible scenarios according to the number of
available transceivers in the physical layer. First there is
considered the case where the cognitive STA is equipped with two
independent transceivers, followed by the more complex case in
which the PHY has only one transceiver available for "over the air"
activities.
Dual-transceiver Mode-PS Mode Disabled
[0099] In this case the cognitive STAs have two independent
transceivers when associated with a PS disabled Basic Service Set
(BSS). In this case a general rule is that each STA establishes one
of its transceivers on the shared common ISM channel in which the
BSS has been established. As a result, this feature aids the
cognitive STAs to more efficiently perform channel sensing
processes. On the other hand, in dual-transceiver systems, the
amount of power consumption is greater than in a single-transceiver
embodiment.
[0100] Beacon frames are assumed to be always transmitted by the
transceiver tuned on the common ISM channel. In addition, managing
the RTS/CTS exchange is performed using this transceiver as well.
For simplicity, one may refer to the transceiver tuned to the ISM
common channel as the ISM transceiver 20, while the other (second)
transceiver, which may be at any frequency band to coordinate
information exchange between cognitive STAs, may be referred to as
the cognitive transceiver 22 (see FIG. 1B).
[0101] The ISM transceiver 20 is responsible for RTS/CTS frames
exchange, and the coordination between cognitive and non-cognitive
STAs. When for the first time a tagged cognitive STA (a "tagged"
cognitive STA is assumed herein to be a considered STA to which the
current focus has been concentrated) desires to establish a
connection with another associated STA in the same BSS, it sends an
RTS frame (comprising the extra fields introduced previously) to
its intended destination. When the received CTS frame corresponding
to the delivered RTS implies that the intended receiver is a
non-cognitive legacy STA, the data exchange is performed in the ISM
band. In this case, as there is no possibility for establishing a
further connection to (with) the tagged STA, for power efficiency
reasons the cognitive transceiver 22 may be turned off (if not
currently involved in any spectrum sensing task). In addition, the
above case is valid when the tagged STA is receiving MSDU(s) from
another non-cognitive STA in the common ISM band. Here, again,
establishing a connection to (with) the tagged STA (another STA) is
not possible since the ISM channel has been already utilized. On
the other hand, when the received CTS frame implies that the
intended receiver is a cognitive STA, the two STAs (i.e., the
cognitive source STA and the cognitive destination STA) can reach a
reciprocal agreement concerning the data channel to be used for
their information exchange. In this case the cognitive transceiver
22 is tuned to the agreed upon channel to commence the exchange of
RTS/CTS frames.
[0102] Assume a case where the tagged cognitive STA is performing
data communication with another cognitive STA in a non-ISM channel
using the cognitive transceiver 22. Now also assume the case where
another STA wants to establish a connection with the tagged
cognitive STA. In this case, when the calling STA is a
non-cognitive STA, the consequent data communication needs to be
done over the shared ISM channel, and there will be no possibility
for other STAs to establish a connection with the tagged STA. On
the other hand, when the calling STA is a cognitive STA, the
concept of non-ISM Band Convergence applies. In this concept, when
a dual-transceiver cognitive STA receives a data exchange request
in the form of an RTS frame from another cognitive STA, while at
the same time being involved with another communication in a
non-ISM channel, it is recommended to invite the calling STA to
also join the utilized non-ISM channel being utilized for the
ongoing data communication. In such cases the Reason Code is used
to convey the tagged STAs intent to the calling cognitive STA. By
using the aforementioned code, the calling STA is enabled to
determine why the called (and tagged STA in the above example) is
inviting it to switch to a certain channel. Generally in this
scenario, the calling STA advertises up to three channels that may
not be the same as those the tagged STAs is interested in. For this
reason the tagged STA is allowed to reject the offered up to three
advertised channels and to advertise another channel instead. To
encourage/force the calling STA to select the tagged STAs desired
channel it may include only its intended channel when it sends a
Negative non-Null CTS back to the calling STA.
[0103] As another use case, consider a cognitive STA that is
communicating with one of its cognitive counterparts in a non-ISM
channel. In addition, consider the scenario when the tagged STA
also wishes to establish an additional connection with another STA,
in parallel with the ongoing communication. The tagged STA uses the
ISM channel to send an invitation message in the form of an RTS
frame to the intended destination STA. When the called STA is not a
cognitive STA, the intended data exchange is performed in the ISM
channel. On the other hand, when the called STA is a cognitive STA,
the communication may be accomplished using any known non-ISM
channel. As above, the concept of non-ISM Band Convergence can be
used, where the tagged STA places its desired (already deployed)
non-ISM channel in the RTS frame to invite the called STA to join.
To encourage/force the called STA to choose the intended channel,
tagged STA may advertise only the aforementioned channel in order
to limit the available options of the called STA. In addition, the
Reason Code is set to convey the reasoning for the advertisement of
the non-ISM channel. By reception of the RTS the called STA can
determine why the tagged STA is promoting this particular non-ISM
channel.
[0104] Note that for both aforesaid cases, it is possible for the
third party (i.e., the called STA) to reject the tagged STA's offer
using the Negative Null control frames (in the first case, using
Negative Null ACK and in the second case, using Negative
Null/non-Null CTS). In this event the third party should disclose
its reasoning regarding the rejection of the offer of the tagged
STA. Using cognitive reasoning, the STAs are able to not only
negotiate with one another concerning channel deployment, but may
also obtain valuable information concerning their surrounding
environment.
Dual-transceiver Mode-PS Mode Enabled
[0105] In this scenario the same definitions regarding the ISM and
cognitive transceivers 20, 22 are applicable, while assuming that
the cognitive STA has two independent transceivers when associated
to a PS-enabled BSS. Basically, in a PS mode enabled network, STAs
contend with each other during ATIM window to inform their intended
receivers about any pending MSDUs.
[0106] In a PS mode enabled BSS, during each ATIM window, those
STAs that have pending MSDU(s) for associated STAs send out ATIM
frames to inform targeted STAs of the presence of buffered MSDU(s)
destined for them. Subsequently, during the remaining time portion
of the Beacon Interval the buffered MSDU(s) are exchanged.
[0107] When the STA has pending MSDU(s) for more than one STA in
the same BSS, it is allowed to contend for the shared wireless
medium as many times as it intends to send ATIM frames to inform
targeted STAs about the existence of buffered MSDU(s); hence, if
the STA is able to contact all targeted STAs in a single ATIM
window, there will be a list of all pending tasks to be
accomplished during the remaining time of the Beacon Interval. In
other words, cognitive STAs are equipped by a "Task List" 24 in
which successive tasks (i.e., acting as either source or
destination STA in different channels) are listed according to
their appearance order and, more importantly, their priority
relative to each other. In this context the words and phrase
"Task", "Duty" and "Upcoming Job" may be used to convey the same
meaning.
[0108] It can be noted that for the dual transceiver mode, each
transceiver 20, 22 may have its own associated Task List 24A, 24B,
meaning that the cognitive transceiver 22 has a dedicated Task List
24A used only for the duties related to non-ISM channels, while the
ISM transceiver 20 has a dedicated Task List 24B for those duties
that may use the shared ISM channel. For the single transceiver
mode all ISM and non-ISM related tasks (i.e., cognitive and
non-cognitive tasks) are maintained in the same (unified) Task
List. 24.
[0109] Within a Task List 24 the task with the highest
priority/position is handled before any other existing/accumulated
task. For example, in FIG. 9 there is a "Reception (1) C.sub.0"task
meaning that upon completion of the current ATIM window the tagged
STA is to switch to channel C.sub.0 to wait for a calling STA
willing to send pending MSDU(s) addressed to it. In this case "(1)"
means that there is only one reception task to be performed in
channel C.sub.0. Note that since the tagged STA knows who will send
pending MSDU(s), upon reception of intended frames from the known
calling STA it simply knows that it should then switch to the next
task. In other words, it is preferred that the tagged STA is not
allowed to switch to the next task before completion of an ongoing
task.
[0110] Continuing with the simple example shown in FIG. 9, as the
second highest priority task the tagged STA switches to channel
C.sub.1 to send a set of pending MSDU(s) to a waiting STA. This
task is shown as "Transmission (1) C.sub.1" meaning that there is
only one transmission task to be done in channel C.sub.1. At the
completion of the second task, the STA switches to channel C.sub.2
to send pending MSDU(s) to another STA. This third task is shown as
"Transmission (1) C.sub.2", meaning that in channel C.sub.2 there
is only one transmission task recorded for the tagged STA.
[0111] As it can be seen from this non-limiting example the task
"Reception (1) C.sub.0" has the highest priority among all of the
listed tasks shown in FIG. 9. In general, "Reception" tasks have a
higher priority than "Transmission" tasks due to the fact that
"Transmission" can be performed at any time during a Beacon
Interval, while "Reception" mandates that the STA switch to the
agreed upon channel and wait for the source STA to complete pending
MSDU(s) reception. As a result, during an ATIM window, if the
cognitive STA is being elected as a receiving STA in a particular
non-ISM channel the "Reception" task is recorded in the non-ISM
(i.e., cognitive) Task List 24A as a duty to be performed with the
highest priority, if there is no other high priority "Reception"
task(s) in the Task List 24A.
[0112] If there is an existing "Reception" task in the Task List,
the new "Reception" tasks may be simply merged with the existing
one. This means that the new "Reception" task, if is going to be
accepted by the tagged STA, should be performed in the same channel
as the existing "Reception" task. This feature may be referred to
as a Channel Convergence Concept and defined for both ISM and
non-ISM channels. There are also other cases for which the same
concept can be used to perform multiple tasks in a single
ISM/non-ISM channel simultaneously.
[0113] As another possible use case, assume that there is an
existing "Reception" task within the Task List 24 with a highest
priority, while the new duty to be added to the Task List 24 is a
task of type "Transmission". The tagged STA prefers to perform the
new "Transmission" task in the same channel as the highest priority
"Reception" duty. This is true at least for the reason that the
cognitive STA prefers to accomplish all "Reception" and
"Transmission" tasks in the same channel to achieve higher channel
throughput by means of traffic aggregation/bursting, while reducing
channel switching energy costs and time. Thus, the STA may offer
the non-ISM channel corresponding to the existing "Reception" duty
to the called STA in order to invite it to join the intended
channel. Now, if the new "Transmission" task is accepted by the
called STA in the offered non-ISM channel, it can be merged with
the existing "Reception" duty and be accomplished in the same
non-ISM channel. This feature follows the Channel Convergence
Concept and is defined for both ISM and non-ISM channels. On the
other hand, if the agreed upon channel corresponding to the new
"Transmission" task differs from the channel used for the existing
"Reception" task, the "Transmission" task is preferably added to
the end of Task List 24 (i.e., with the lowest priority).
[0114] When a tagged STA sends/receives an ATIM frame, it is
involved in an interaction in which both parties (i.e., the tagged
STA and the other party) are allowed to negotiate concerning their
desired channel(s) to be used during their data communication. When
the calling STA begins the interaction, it provides sufficient
reasoning regarding its chosen non-ISM channel(s). This reasoning
is done with the help of Task List 24 and the pre-defined Reason
Codes discussed above. As was explained, each STA has a set of
tasks listed according to their order of appearance, importance,
and priorities. Among all existing tasks, there is a task with the
highest priority that is to be used by the STA. This task may be a
mixed task, e.g., "Reception (3) C.sub.0" which means that exactly
three MSDU receptions are to be performed in channel C.sub.0 from
three different source STAs. Using the highest priority task, the
calling STA begins the interaction and includes its own reasoning
based on the highest priority task within the delivered ATIM frame.
On the other side the called STA receives the ATIM frame and, based
on the provided Reason Code within the ATIM frame, the called STA
reacts appropriately according to a set of pre-defined mandatory
rules. These pre-defined rules, on one hand, take essential key
points for optimized multiple data communications into account
while, on the other hand, prevent both involved parties from
acting/reacting in a self-interested only manner. In general, the
interacting parties (involved STAs) are concerned with their
highest priority tasks when offering their intended channels to one
another and, based on the existing rules, they reach a mutual
agreement concerning a data channel.
[0115] Note that if the task with the highest priority is a merged
"Reception" task with a "Transmission" duty in the same channel,
then preferably only the "Reception" task is used for channel
negotiation.
[0116] The pre-defined rules are now explained in further detail.
Recall that the interacting parties are concerned with their
highest priority task at hand, and the "winner" in a particular
interaction is a STA for which the offered channel is accepted by
the other STA. Upon determination of the winner, there is another
set of rules which define how the new assigned task should be
recorded in the Task List 24 (i.e., should it be merged with the
existing high priority task or simply added to the list with the
lowest priority).
[0117] Consider first the four exemplary modes, shown in FIGS.
10A-10D, which will be used below.
[0118] Assume that, for the first time, the tagged STA to be
informed by another cognitive STA of a pending MSDU that is to be
sent to the tagged STA. After the channel negotiation phase the
STAs reach a mutual agreement to use a non-ISM channel after the
completion of the ongoing ATIM window. As a second event, consider
the case where the tagged STA is once more informed of a pending
MSDU by another cognitive STA. The concept of the non-ISM Band
Convergence should thus be used. The tagged STA receives an ATIM
frame advertising a set of non-ISM channels and possibly none of
the advertised channels are the same as the desired channel of the
tagged STA (which is in fact the one that has been agreed to
previously in the former ATIM channel negotiation phase). In this
case two different scenarios are possible.
[0119] A1. When the Reason Code in the received ATIM frame is 000,
the tagged STA rejects the offered channel in the ATIM frame and
sends a Negative non-Null ACK back to the calling cognitive STA,
and offers the channel preferred by the tagged STA. In this case
the calling STA should accept the offer by the tagged STA if the
advertised non-ISM channel is not in its local Primary User
Appearance (PUA) table 30 (discussed below with reference to FIG.
15), and if it has satisfactory spectrum quality results. Upon
completion of ongoing ATIM window, the calling cognitive STA first
sends the pending MSDU(s) in its own channel, and then switches to
the channel of the tagged STA channel and sends the MSDU(s)
addressed to the tagged STA (Mode II).
[0120] B1. When the Reason Code in the received ATIM frame is 001,
the tagged STA rejects the offer in the ATIM frame and sends a
Negative non-Null ACK back to the calling cognitive STA, and offers
the channel preferred by the tagged STA. In this case the calling
STA should accept the offered channel by the tagged STA if the
advertised non-ISM channel is not in its local PUA table 30, and if
it has satisfactory spectrum quality results. Upon completion of
ongoing ATIM window, the calling cognitive STA waits for its
pending MSDU(s) in its own channel and, after reception of the
intended MSDU(s), it switches to the channel of the tagged STA and
sends the MSDU(s) addressed to the tagged STA (Mode III).
[0121] Assume now that instead of a cognitive STA, a legacy
non-cognitive STA informs the tagged STA about a pending MSDU.
Since the tagged STA is equipped with the dual transceiver 20, 22
system, it is able to accept this request to let the calling legacy
STA send the pending MSDU(s) after completion of the ongoing ATIM
window (over the ISM channel associated with ISM transceiver
20).
[0122] Alternatively, again assume that after completion of an
ongoing ATIM window the tagged STA should be a receiving STA in a
non-ISM channel, but subsequently wants to send a pending MSDU to
another associated STA in the wireless network. In this case the
tagged STA is allowed to send an ATIM frame to inform the intended
receiving STA. However, since prior to any ATIM transmission the
tagged STA has no idea whether its intended receiver STA is a
cognitive STA or a legacy, non-cognitive STA, it preferably assumes
that the receiver is a cognitive STA and puts the formerly agreed
upon non-ISM channel in the ATIM frame. In addition, the tagged STA
is mandated to set the Reason Code of the ATIM frame to 001. In
this case two different scenarios are possible, if one assumes that
the called STA is a cognitive STA.
[0123] A2. When the called cognitive STA was to be a receiving STA
in the following Beacon Interval, and the offered channel is not
the same as its desired non-ISM channel, it rejects the channel
offered by the tagged STA and responds with a Negative non-Null ACK
containing its own desired channel. The tagged STA should then
accept the offer if the channel it is not in its local PUA table
30, and if the offered channel has satisfactory spectrum quality
results. Upon completion of the ongoing ATIM window, the tagged STA
waits for its pending MSDU(s) in its channel and, after receiving
the intended MSDU(s), it switches to the channel of the called
cognitive STA and sends the pending MSDU(s) addressed to the called
STA (Mode III).
[0124] B2. When the called cognitive STA was to be a source STA in
the following Beacon Interval, and the offered channel is not the
same as its desired non-ISM channel, it accepts the channel offered
by the tagged STA if it is not in its local PUA table 30, and if it
has satisfactory spectrum quality results. Upon completion of the
ongoing ATIM window the called STA first switches to the desired
channel of the tagged STA and waits to receive its intended
MSDU(s), then upon reception of the pending MSDU(s) it switches
back to its own desired channel and sends the pending MSDU(s) to
its intended receiver(s) (Mode IV).
[0125] Assume now that instead of a cognitive STA, a legacy
non-cognitive STA is informed by the tagged STA about a pending
MSDU. Since the tagged STA is equipped with the dual transceiver
20, 22 system, it is able to handle MSDU delivery after completion
of the ongoing ATIM window (over the ISM channel associated with
ISM transceiver 20).
[0126] It should be noted that if, for the first time, the tagged
STA is informed by a legacy non-cognitive STA about a pending MSDU,
it is allowed to accept any other cognitive/non-cognitive ATIM
request without any particular limitation. The accepted legacy ATIM
requests are coordinated over the ISM channel by the use of ISM
transceiver 20.
[0127] Assume now that, as in the first case, the tagged STA
intends to transmit a pending MSDU to another STA after completion
of the current ATIM window. Also assume that the transmission is
going to be made to a non-cognitive STA over the shared ISM
channel. This transmission can be accomplished without confusion
with any non-ISM data transmission during the upcoming Beacon
Interval. At this point, and as the second event, consider the
following cases.
[0128] A3. In a first case the tagged STA is informed about a
pending MSDU by another cognitive STA. In response, the calling and
tagged cognitive STAs negotiate to reach a mutual agreement
concerning a non-ISM channel, and subsequently the tagged STA
cognitive transceiver 22 is tuned to the agreed upon non-ISM
channel. Afterwards, if a pending MSDU transmission request in the
form of an ATIM frame is received from any other cognitive STA, the
concept of non-ISM Band Convergence is utilized according to the
rules A1, B1 discussed above. When the tagged STA receives any
other requests from non-cognitive STAs, it simply handles them in
the ISM channel in a manner similar to the IEEE 802.11 legacy PS
mode MAC. However, when the tagged STA desires to send any ATIM
frame to any other cognitive STA, the concept of non-ISM Band
Convergence is utilized according to the rules A2, B2 discussed
above.
[0129] B3. Another case is concerned with when the tagged STA is
informed about a pending MSDU by another non-cognitive STA. This is
exactly the same situation as the IEEE 802.11 legacy PS mode MAC,
meaning that the STA handles all requests (all together) in the ISM
channel. In meantime, when the tagged STA receives an invitation
from a cognitive STA, the former case is followed.
[0130] C3. Another case is concerned with when the tagged STA
desires to inform another associated STA about a pending MSDU.
Since, at this moment, the tagged STA has not agreed to use any
particular non-ISM channel, it is allowed to advertise up to three
channels based on its localized decision making/reasoning. If the
called STA is a non-cognitive STA, it shall respond with an ACK
frame and subsequently the transmission will be performed over the
ISM channel. On the other hand, if the called STA is a cognitive
STA, the two STAs shall negotiate a non-ISM channel to be used.
From this point on, all possible events are addressed according to
the rules A4, B4, A5, B5 discussed below.
[0131] Now assume that as the first possible case the tagged STA
was already supposed to transmit a pending MSDU to another STA
after completion of the current ATIM window. Also assume that the
aforementioned transmission is going to be made to a cognitive STA
in an agreed upon non-ISM channel. At this point, and as a second
event, consider the case where the tagged STA is informed about a
pending MSDU by another cognitive STA. In this case two different
scenarios are possible.
[0132] A4. When the Reason Code in the received ATIM frame is 000,
the tagged STA accepts the offer if it is not in its local PUA
table 30, and it has obtained satisfactory spectrum quality
results. Upon completion of ongoing ATIM window, the tagged STA
switches to the offered channel to receive its pending MSDU(s) from
the calling STA, then it comes back to its channel and sends the
pending MSDU(s) to its intended receiver(s). This feature enables
the calling cognitive STA to enhance channel utilization by using a
bursting scheme. On the other hand, when frame bursting is not
possible the calling STA sends the MSDU(s) of the tagged STA before
any other pending MSDU (Mode I).
[0133] B4. When the Reason Code in the received ATIM frame is 001,
the tagged STA accepts the offer if it is not in its local PUA
table 30, and if it has obtained satisfactory spectrum quality
results. Upon completion of the ongoing ATIM window the tagged STA
switches to the offered channel to receive its pending MSDU(s) from
the calling STA, and then it switches back to its channel and sends
the pending MSDU(s) to its intended receiver(s) (Mode IV).
[0134] Assume now that instead of a cognitive STA, a legacy
non-cognitive STA informs the tagged STA of a pending MSDU. Since
the tagged STA is equipped with the dual-transceiver 20, 22 system,
it is able to accept this request to let the calling legacy STA
send the pending MSDU(s) after completion of the ongoing ATIM
window (over the ISM channel)
[0135] Alternatively, again assume that after completion of an
ongoing ATIM window the tagged STA is to be a source STA in an
agreed non-ISM channel, but subsequently it desires to send a
pending MSDU to another associated STA in the wireless network. In
this case the tagged STA is allowed to send an ATIM frame to inform
the intended receiving STA. However, since prior to any ATIM
transmission the tagged STA has no idea as to whether its intended
receiver is a cognitive STA, it preferably assumes that the
intended receiver is a cognitive STA and it places the formerly
agreed upon non-ISM channel in the ATIM frame. In addition, the
tagged STA is also mandated to set the Reason Code of the ATIM
frame to 001. In this case two different scenarios are possible, if
it is assumed that the called STA is a cognitive STA.
[0136] A5. When the called cognitive STA was to be a receiving STA
in the following Beacon Interval, and the offered channel is not
the same as its desired non-ISM channel, it rejects the offer of
the tagged STA and responds with a Negative non-Null ACK containing
its own desired channel. The tagged STA accepts the offer if it is
not in its local PUA table 30, and if it has obtained satisfactory
spectrum quality results. Upon completion of the ongoing ATIM
window, the tagged STA first sends the pending MSDU(s) in its own
desired channel to its intended receiver(s), then it switches to
the channel of the called STA channel and sends the pending MSDU(s)
to it (Mode II).
[0137] B5. When the called cognitive STA was to be a source STA in
the following Beacon Interval, and the offered channel is not the
same as its desired non-ISM channel, it accepts the offer of the
tagged STA if it is not in its local PUA table 30, and if it has
obtained satisfactory spectrum quality results. Upon completion of
the ongoing ATIM window the called STA first switches to the
desired channel of the tagged STA and waits for it to send its
intended MSDU(s), then upon reception of the pending MSDU(s), it
switches back to its own desired channel and sends pending MSDU(s)
to its intended receiver(s). This feature enables the tagged STA to
enhance channel utilization by using bursting. On the other hand,
when frame bursting is not possible, the tagged STA sends the
MSDU(s) of the called STA before any other pending MSDU (Mode
I).
[0138] Assume now that instead of a cognitive STA, a legacy
non-cognitive STA is informed by the tagged STA of a pending MSDU.
Since the tagged STA is equipped with the dual-transceiver 20, 22
system, it handles the MSDU delivery after completion of the
ongoing ATIM window over the ISM channel and by the use of its ISM
transceiver 20.
[0139] After determining the winner of interaction (negotiation),
both involved STAs update their Task Lists 24 appropriately. FIGS.
11 and 12 show how the new task is added to the Task List 24
according to the different modes, explained earlier, and the
position of party (whether it was a calling STA, i.e., Side A, or a
called STA, i.e., Side B).
[0140] Explained now in greater detail is how the above-described
rules may be used when merging a new task with the existing Task
List 24. Consider the case when the Task List 24 has a "Reception
(1) C.sub.0" at the top (highest priority) of its recorded tasks
(FIG. 11). As in the first row in FIG. 11A, assume that the owner
of Task List 24 is invited to participate in Mode III as the called
STA (i.e., Side B). According to the rules explained earlier, the
tagged STA becomes the winner and as a result the calling STA
accepts its offer, i.e., accepts C.sub.0. This means that there
will be a new task for the called STA as "Reception", for which the
agreed upon channel is exactly the same as the existing high
priority task in the Task List 24. Therefore, the new task can be
simply merged with the existing task (Convergence Concept) and,
subsequently, the Task List 24 becomes similar to that illustrated
list in FIG. 11. In the third row, since the new task is
"Transmission (1) C.sub.1", it has the lowest priority and as a
result is recorded according to its appearance order. Hence, as
shown in FIG. 11 it is put at the end of the Task List 24. On the
other hand, in the fourth row, since the new "Transmission" task
has exactly the same agreed upon channel as the existing
"Reception" task, it is simply merged with it and, consequently, is
be handled in parallel with the "Reception (1) C.sub.0" upon
completion of the ATIM window.
[0141] Similarly, FIG. 12 illustrates all possibilities when a
"Transmission" task is at the top of the Task List 24.
[0142] During the Beacon Interval, when two cognitive STAs desire
to establish a layer two connection, they preferably use RTS/CTS
frames to negotiate the channel to be used later for the data
exchange. When the intended receiver is unreachable for any reason
(e.g., it is situated in the radio range of an ongoing large data
communication in the calling STA's advertised channel), it is
allowed to reject the offered channel and advertise its own
suggestion using the Negative non-Null CTS , with the Reason Code
(see FIG. 6) set to 110. Upon reception of the Negative non-Null
CTS, the source STA checks the Reason Code and is able to determine
the reason for which its desired channel has been rejected by the
receiver. At this point, according to the source STA decision
making it decides whether to accept the channel counteroffer of the
receiver.
[0143] All Beacon frames are sent simultaneously over the common
ISM channel utilized by legacy and cognitive STAs. Beacon frame
generation is performed in a cooperative fashion by all associated
cognitive STAs, in the same manner as it is performed by legacy
STAs and as defined in the IEEE 802.11 standard. During Beacon
frame delivery all STAs become quiet and all information exchange
activities are suspended. Cognitive STAs are encouraged to put
information about Recently Discovered Channels in generated Beacon
frames. On the other hand, cognitive STAs are mandated to put
sufficient information concerning those channels in which a Primary
User has been detected (appeared), accompanied by the corresponding
Time Index. It is imperative to note that the encapsulated
information regarding the Primary User Appearance is not
necessarily based on local induction (interpretation), but may be
obtained from the received Beacon, Probe response, or other types
of MAC management frames captured over the air interface.
Furthermore, it should be also pointed out that for Recently
Discovered Channels there should not be any accompanying Time Index
within the Beacon or other management frames (in this case the
Probe Response management frame is not used for Recently Discovered
Channel announcement). As it will be clarified below, the cognitive
STA examines each discovered channel to conclude whether it should
be recorded in local databases. When the cognitive STA experiences
a non-ISM channel successfully, either in the position of a source
STA or in the position of a destination STA, the channel is
recorded in a so-called Last Successfully Experienced Channel
(LSEC) database (simply LSEC table 32, shown in FIG. 15 and
discussed below). Only when cognitive STA's local interpretation
implies that utilization of a particular non-ISM channel provides a
minimum level of satisfaction, the channel information can be
recorded in the local LSEC table 32. Note that currently the
defined level of satisfaction covers only an ability of successful
experience, and an acceptable spectrum quality result.
[0144] The Time Index is specified in a number of Beacon Intervals
elapsed since an event has taken place.
[0145] As was discussed above, the MAC entity 26 uses the LSEC
table 32 of those channels that have been successfully experienced
by the tagged STA. Each entry of the LSEC table 32 has two fields:
the first field is a Channel Identifier (e.g., center frequency,
channel number, etc.) and the second field is its corresponding
Time Index (the number of Beacon Intervals that have elapsed since
the last successful experience with the channel by the tagged STA).
The Time Index of LSEC entries is a monotonically increasing timer
that increments every Beacon Interval. When the tagged STA utilizes
one of the recorded channels in the LSEC table 32 successfully,
either as a source or a destination end point, the corresponding
Time Index is reset to zero. By definition, a recorded channel in
the LSEC table 32 is referred to as an LSEChannel. To keep track of
the recorded information level of freshness, the Time Index Domain
is divided to two sub-domains (or zones) as depicted in FIG.
13.
[0146] The first zone (sub-domain), LSEChannel Forbidden Zone 29A,
is typically spanned from 0 to M'. When a discovered channel is
successfully experienced by the tagged STA (either as a source or
destination STA), it receives a Time Index equal to zero and is
recorded in the LSEC table 32. The second zone, LSEChannel Election
Zone 29B, is spanned between M'+1 and M. As explained earlier, when
a previously recorded channel in the LSEC table 32 is again
experienced successfully, its Time Index is reset to zero.
Basically, the MAC entity 26 always selects a non-ISM channel with
the greatest Time Index as its desired choice for an upcoming data
communication. This feature makes the non-ISM Band Divergence
Concept achievable, as it prevents greedy utilization of a
particular non-ISM channel by cognitive STAs after any successful
experience with the channel. On the other hand, this feature is
indeed imperative for the case where a Primary User appears in a
certain non-ISM channel. If all cognitive STAs concentrate on
deployment of only one non-ISM channel, upon any Primary User
Appearance they should concurrently suspend all ongoing information
exchanges, resulting in simultaneous and considerable system
performance degradation. If associated cognitive STAs are
distributed among all available non-ISM channels (i.e., frequency
opportunities), then upon a Primary User Appearance in a particular
channel only a few cognitive STAs will encounter throughput
degradation due to traffic exchange suspension. Note that the MAC
entity 26 is not allowed to update the LSEC record Time Index with
any type of received information over the air interface. In
addition, upon reporting a new LSEChannel, the reporter is
prevented from inclusion of any Time Index in the management frame
(e.g., Beacon frame) carrying the LSEChannel information.
[0147] In addition to the LSEC table 32, there is a Primary User
Appearance (PUA) table 30 used to keep track of all channels in
which a Primary User has been recently detected. When a Primary
User is observed in a channel, using an extended version of a
Dynamic Frequency Selection (DFS) scheme (refer to IEEE 802.11h
amendment), cognitive STAs are mandated to inform their one-hop
neighbors about Primary User appearance in that particular channel.
All discovered channels, in which Primary User appearance has been
reported by either the local MAC entity 26 or neighboring STAs,
should be recorded in PUA table (Note that the terms PUA database
and PUA table can be used interchangeably). In PUA table, each
entry has a Time Index which is a monotonically decreasing timer
and decrements every Beacon Interval. When a channel is being
recorded in the PUA table 30, its timer is loaded by a pre-defined
value (e.g., a network operator-defined value) and begins counting
down. When the Time Index timer reaches zero, the channel
identifier is simply removed from the PUA table 30. When a
Beacon/Probe Response/DFS frame reporting a Primary User Appearance
is received, the cognitive STA checks its private PUA table 30 to
verify whether the reported channel has been already recorded in
its database; if not, it is recorded and its corresponding Time
Index is simply copied from the received frame. Note that in
contrast to the "Recently Discovered Channel" case where the
reporter is prohibited from inclusion of the Time Index in the
delivered management frames (e.g., Beacon frame), for the case of
"Primary User Announcement" the reporter is mandated to accompany
the channel being reported with its corresponding Time Index. On
the other hand, if the Primary User Appearance has already been
recorded in PUA table 30, the corresponding Time Indexes in the
local database and the received report frame are compared with each
other. When the Time Index in PUA table 30 is less than the Time
Index in the received report frame, the cognitive STA updates the
Time Index of the channel entry in its local database with the Time
Index included in the received frame. When the Time Index in PUA
table 30 is greater than the Time Index in the received frame, the
cognitive STA simply drops the received information; i.e.:
TABLE-US-00001 If (Rcvd_Frame(Channel_Id(Time_Index)) .ltoreq.
Local_PUA_Table(Channel_Id(Time_Index))) { Drop the information
inside the received frame; } Else {
Local_PUA_Table(Channel_Id(Time_Index)) .rarw.
Rcvd_Frame(Channel_Id(Time_Index)); }
[0148] When a Primary User is detected in an LSEChannel, the
channel should be removed from the LSEC table. As a result, upon
receiving a Primary User Appearance notification reported either by
the local MAC entity 26 or by a neighboring STA, not only the PUA
table 30, but also the LSEC table 32, should be checked. All
dynamic frequency selection (DFS) related frames are transmitted
over the ISM channel.
[0149] Within, for example, the Beacon and Probe Response frames
there should be an Information Element dedicated to Primary User
Appearance notification. This field resides in the optional portion
of above mentioned management frames. Note that the corresponding
Time Index for Primary User Appearance event should also be
included. Thus, the Channel Identifier (e.g., center frequency) and
Time Index are two mandatory sub-fields that are always
included.
[0150] Reference can be made to FIG. 14 for showing this
information element for Primary User Appearance Reporting
purposes.
[0151] In addition to the LSEC table 32 and the PUA table 30 there
is a section within the MAC entity 26 databases referred to as a
Transient Zone 34 that is specifically used for channel management
purposes (See FIG. 15). All information concerning discovered
channels either by the local MAC entity 26 or neighboring STAs is
gathered within the Transient Zone 34. A newly discovered non-ISM
channel is first checked (block 36) to see whether is already
included in the PUA table 30. After checking the PUA Table 30, and
if the verified channel has not been recorded in the PUA table 30,
the channel information is recorded in a FIFO-based structure 38.
The FIFO-based structure 38 is actually a FIFO queue, and channel
information is queued in this structure and fetched according to
inclusion order. On the other hand, if the discovered channel is
found in the PUA table 30, it is simply discarded from the
Transient Zone 34.
[0152] Head-of-line (HOL) channel information is fetched from the
Transient Zone FIFO-based queue 38 whenever there is no channel in
the LSEC table 32 with a Time Index situated in the LSEChannel
Election Zone 29B (See FIG. 13). Basically, as a first choice, the
MAC entity 26 is mandated to select a candidate channel with a
largest Time Index situated in LSEChannel Election Zone 29B. When
there is no channel with the Time Index situated in LSEChannel
Election Zone 29B, the MAC entity 26 should check the Transient
Zone 34 for any channel availability. When at least one available
channel is found in the Transient Zone 34, it is used for medium
access and data exchange purposes. If the deployed channel results
to a successful data communication, it is recorded in the LSEC
table 32 accompanied by a Time Index equal to zero; otherwise, the
channel information is sent back to the FIFO queue 38 as
illustrated in FIG. 15. Finally, when there is no channel with a
Time Index situated in LSEChannel Election Zone 29B, and the
Transient Zone is empty (FIFO Queue 38 is empty), the MAC entity 26
should select a channel with a largest Time Index in the LSEChannel
Election Forbidden Zone 29A (or simply, LSEChannel Forbidden Zone)
shown in FIG. 13.
[0153] If a channel selected from the LSEC table 32 is utilized for
a data communication successfully, its corresponding Time Index
will be reset to zero and it will be retained in the LSEC table 32.
On the other hand, if the aforementioned channel is utilized for an
unsuccessful data communication, it will be put into the Transient
Zone 34 as illustrated in FIG. 15.
[0154] If a channel selected from the Transient Zone 34 is utilized
for a data communication successfully, it will be recorded in the
LSEC table 32 and its corresponding Time Index will be set to zero.
On the other hand, if the abovementioned channel is utilized for an
unsuccessful data communication, it will be returned to the
Transient Zone 34 as shown in FIG. 15.
[0155] To select a channel for an upcoming communication the MAC
entity 26 first inspects the LSEC table 32 to determine whether a
channel with a largest Time Index situated in LSEChannel Election
Zone 29B can be found. When there is at least one such channel, it
is selected to be advertized during the channel negotiation phase.
Otherwise, the MAC entity 26 inspects the Transient Zone 34 and
determines if channel information exists in the FIFO queue 38; if
at least one channel is found, it is selected. If not, the MAC
entity 26 reverts to the LSEC table 32 and determines if any
channel with a largest Time Index situated in LSEChannel Forbidden
Zone 29A. To review, first the LSEC table 32 is checked for any
available channel with the largest Time Index situated in the
LSEChannel Election Zone 29B , then the Transient Zone is checked,
and finally if the MAC entity 26 needs to choose additional
channel(s) it checks the LSEC table 32 for any available channel
with a largest Time Index in LSEChannel Forbidden Zone 29A.
[0156] The aforementioned structure makes the use of the non-ISM
Band Divergence Concept possible by preventing greedy utilization
of a particular non-ISM channel by cognitive STAs after a
successful experience with the channel. In other words, the MAC
entity 26 operates to avoid accumulated channel requests for a
certain non-ISM channel. Generally, when a successfully experienced
channel is time indexed by zero, the MAC entity 26 is prohibited
from deploying it again and, as a result, it becomes available for
use by other neighboring STAs. The use of this approach results in
reduced channel congestion levels and to a reduced backoff delay
for highly populated wireless networks. As was noted above, the use
of this approach is also important for the case where the Primary
User appears in a certain non-ISM channel. If all associated
cognitive STAs 10A have already concentrated in only one non-ISM
channel, then upon a Primary User Appearance in this particular
channel all STAs 10 would have been mandated to concurrently
suspend all ongoing information exchanges. In contradistinction, if
the cognitive STAs 10A have been distributed amongst all available
non-ISM channels (frequency opportunities), then upon Primary User
Appearance in a particular channel only some subset of the set of
all cognitive STAs would encounter throughput degradation due to
traffic exchange suspension.
[0157] The local MAC entity 26 is preferably capable of discovering
new frequency opportunities in a continuous fashion. In addition,
the tagged STA is able to discover new channels from its neighbors
by simply listening to their exchanged control frames (e.g.,
ATIM/RTS/CTS/ACK/Negative non-Null CTS/Negative non-Null ACK). By
using this strategy the STAs 10 share their findings in a
cooperative and reactive manner.
[0158] As was explained previously, the cognitive STAs 10A are
mandated to utilize the ISM channel for BSS establishment,
management, and control information exchange. On the other hand,
according to the IEEE 802.11 standard if a legacy STA 10B receives
a management/control frame accompanied with one of the known Frame
Types defined in IEEE 802.11 it processes only those fields that it
knows how to exploit. This means that newly added fields are not
processed and will be simply discarded by the legacy STAs 10B. On
the other hand, the legacy STAs 10B should be able to process the
Duration/ID field and defer from medium access during the period
when two cognitive STAs 10A are exchanging control frames. For this
reason it is desirable to define how the Duration/ID field should
be set when exchanging control frames over the shared ISM
channel.
[0159] The source STA 10 is allowed to advertise up to three
channels within an ATIM/RTS frame. As a result, generally ATIM/RTS
frames have no static pre-defined size as in legacy IEEE 802.11.
Despite the dynamic nature of the ATIM/RTS frame size, no
complication arises due to the fact that the source STA 10 may only
acquire the control of the medium, and subsequently is free to
transmit its ATIM/RTS without any limitation on the size of
delivered frame. However, at the receiver side a problem may arise
since it may not be clear how the receiver will respond to the
received ATIM/RTS frame. As a result, the source STA 10 does not
know exactly for how much time it should reserve its surrounding
wireless medium, using the network allocation vector (NAV) within
the ATIM/RTS frame, to enable the called STA to send the expected
ACK/CTS frame back. For example, in one case it is possible that
one of the offered channels is accepted by the receiver and,
consequently, only CC fields will be included within the returned
ACK/CTS frame. In this case the source STA 10 can readily predict
that if the receiver is going to accept the offer, it will send
back three CC fields (one for the accepted channel and two for the
rejected channels). However, there may instead be a case where the
receiver rejects all offered channels, and also advertises up to
three channels of its own in the Negative non-Null ACK/Negative
non-Null CTS frame. As should be apparent, in general one may
conclude that there is no way for both the source and destination
STAs 10 to predict how the other end-point may respond.
[0160] To solve this problem it is preferred that the source STA 10
reserve the medium to accommodate a maximum possible returned frame
size. As a result, and in accordance with the exemplary
embodiments, the RTS reserves the medium for a CTS frame containing
up to three CC fields (the source STA 10 knows exactly how many CC
fields will be included in a returned CTS since they exactly
correspond to the number of channels previously offered in the
RTS), in addition to as many as three CI fields (corresponding to
the maximum possible three channels that the receiver is allowed to
advertise, if it rejects the three channels offered by the source
STA 10). As an example, assume that the source STA 10 advertises
two non-ISM channels, and that the receiver does not accept the
offered channels and prefers instead to advertise only one non-ISM
channel different from those offered by the source STA. According
to the aforementioned rule, the source STA reserves its surrounding
area for:
SIFS+legacy_CTSTime+2.times.CC+3.times.Cl.
[0161] This means that source STA 10, due to its inability to
predict how many channels may be offered by its intended receiver,
reserves the medium for the case when the receiver offers three
different channels.
[0162] Related to the dynamic size of the CTS there is a more
conservative approach, that is, for the receiver limit the allowed
number of offered channels to one. In this case the source STA is
able to anticipate that if all offered channels are rejected by its
intended receiver then there will be exactly the same number of CC
fields as the number of offered channels in the RTS, plus one CI
field, in the returned CTS frame (i.e., Negative non-Null CTS).
Although in this case the size of returned Negative non-Null CTS
frame is still unpredictable, the incurred variation is not
considerable in comparison to the case in which receiver STA is
allowed to offer up to three channels. Reconsidering the previous
example, according to the above simplifying rule, the source STA is
only required to reserve the channel for:
SIFS+legacy_CTSTime+2.times.CC+Cl.
[0163] On the other hand, if the receiver STA 10 has no channel
offer to advertise, the additional media reservation performed by
the source STA 10 will be equivalent to one CI, which is only five
bytes in size and is thus not a significant portion of the overall
reservation.
[0164] For the case of the receiver STA, when it reserves its
surrounding media the situation is simpler. If the receiver STA
accepts one of the offered channels, it will not need to reserve
the media for any additional period. On the other hand, if it
rejects all offers and advertises its own desired channel, it can
anticipate an ACK (or Negative ACK) frame with exactly the same
number of CC fields as the number of offered channels in its
Negative non-Null CTS.
[0165] As can be appreciated, the media over-reservation potential
problem exists only for the source STA 10, and by limiting the
number of the receiver STA possible offers the problem can be at
least partly alleviated.
[0166] The following rules are applicable for involved STAs 10
concerning reservation of the wireless medium:
[0167] 1. ATIM/RTS frame Duration/ID field should be loaded by
SIFS+legacy_CTS Time+number of offered channels by source
STA.times.CC+maximum number of channels receiver is allowed to
advertise if it is going to reject all source STA's offers.times.CI
(over-reservation is probable).
[0168] 2. Positive CTS/ACK and Negative Null CTS/ACK frames
Duration/ID is loaded by zero since there is no need to reserve the
media for extra frame reception.
[0169] 3. Negative non-Null CTS/ACK frame Duration/ID should be
loaded by SIFS+legacy_ACK Time+number of offered channels by
receiver.times.CC (no over-reservation).
[0170] Discussed thus far has been the multiple transceiver
cognitive STA 10A embodiment, having the ISM transceiver 20 and the
cognitive transceiver 22. Discussed now is the single transceiver
embodiment.
[0171] When the physical layer is equipped with a single
transceiver, several issues arise concerning multi-channel MSDU
transmission. In this discussion it may be assumed that the
cognitive STAs 10 are single-transceiver-based and as a result,
each cognitive STA 10A is able to focus on only one channel, either
ISM or non-ISM channel, at any given time. Therefore, for example,
if the STA 10A is sending or receiving on a non-ISM channel, then
it cannot simultaneously handle legacy ISM connections. Both the
power saving (PS) mode disabled and the PS mode enabled scenarios
are covered below.
A) PS Mode Disabled
[0172] This mode is similar to the case where the physical layer is
equipped with a dual transceiver (ISM/non-ISM), except that the
cognitive STA 10A is not able to serve any connection establishment
request delivered over the shared ISM channel when it is involved
with an ongoing data transmission on a non-ISM channel. This is due
to the fact that the STA has the single transceiver structure and,
thus, it can be tuned to only one channel at any given time. As a
result the STA which is sending an ATIM/RTS frame to establish a
connection with a cognitive STA 10A that has being involved with
another data communication in a non-ISM channel will not receive
any response and subsequently will timeout. The occurrence of
successive timeouts degrades the ISM channel utilization and wastes
available opportunities for other legacy STAs 10B to establish data
communication with their associated counterparts. It can be assumed
that the achieved channel utilization in single-transceiver mode is
lower than for the case in which the STA is equipped with two
independent transceivers. Also, it should be noted that the
aforementioned throughput degradation is due to the overall
structure of physical layer, and is not due to the MAC entity 26
itself.
B) PS Mode Enabled
[0173] In this case all transmitted/received MSDU(s) delivery
requests, in the form of ATIM, are handled exactly as in the dual
transceiver mode, except that all ISM and non-ISM related
"Transmission" and "Reception" tasks that are to be performed in
the upcoming Beacon Interval are accumulated in a common Task List
24. This means that there is no separate non-ISM and ISM Task Lists
24A, 24B, due to the fact that a single transceiver-based STA 10 is
capable of handling only one task at any particular time.
Therefore, successive tasks are recorded according to their
appearance order and, more importantly, based on their priorities.
In conclusion, for the single transceiver case there is only one
Task List 24 dedicated to both ISM and non-ISM duties.
[0174] Concerning the strategic interactions between cognitive STAs
10A during the negotiation phase, all of the rules discussed above
are applicable with the following exception.
[0175] When there is at least one "non-ISM Reception" task in the
(unified) Task List 24, any "ISM Reception" task issued by a legacy
STA 10B is rejected. On the other hand, when there is an "ISM
Reception" task in the Task List 24, all other "Reception" tasks
are converged to the common ISM channel, as the Convergence Concept
has been taken into account for all pre-defined interaction rules
introduced earlier. This may be referred to as `ISM Attraction
Dilemma`. The ISM Attraction Dilemma is undesirable in the sense
that it attracts all cognitive transmissions to the shared common
ISM band, which preferably is used only for control signaling
purposes. Therefore, the "ISM Reception" task(s) are preferably
rejected if there is at least one "non-ISM Reception" task in the
Task List 24. Since legacy STAs 10B are not capable of switching to
non-ISM channels, as a result they are not able to follow the
Convergence Concept in non-ISM channels and, for this reason, their
transmission requests over the ISM channel are simply rejected (if
the cognitive STA 10A has already agreed to a "Reception" task over
a non-ISM channel).
[0176] FIG. 16 shows a modified Task List maintenance rule and
strategy that is based on the foregoing discussion of the single
transceiver embodiment of the STA 10.
[0177] As it can be seen in FIG. 16, the ISM Attraction Dilemma is
possible for three different cases. Since the tagged STA has an
"ISM Reception" task, as a result all other accepted "Reception"
tasks are also attracted to the ISM channel. For the fourth row,
there is also a possibility of ISM Attraction, if the existing
"Reception" task in Task List 24 is related to a legacy STA
10B.
[0178] Based on the foregoing it can be appreciated that a novel
cognitive frequency agile MAC protocol entity 26 has been
described, one suitable for use with, but not limited to, next
generation 802.11 wireless networks. The MAC entity 26 is capable
of coordination of concurrent multi-channel data communications in
a distributed fashion. Both the power save mode disabled and power
save mode enabled modes are accommodated while taking into account
both single transceiver and dual transceiver physical layer
structures. The MAC entity 26 protocol algorithm is designed in
such a way that it makes both non-ISM channel Convergence and
Divergence Concepts simultaneously possible and achievable. The use
of these two concepts improves channel utilization on one hand, and
on the other hand prevents congestion, in particular in non-ISM
channels. For each of aforementioned concepts novel strategies,
including the use of dual-zone timers, strategic decision-making
schemes and the Task Lists 24A, 24B, have been described. In
addition, it can be noted that the exemplary embodiments do not
require any dedicated control channel structure for the cognitive
STAs 10A, while providing additional services for legacy existing
IEEE 802.11 networks. Also, the cognitive MAC entity 26 has
complete backwards compatibility with the legacy IEEE 802.11 MAC,
meaning that the legacy STAs 10B are able to receive and process
all MAC frames transmitted by the cognitive STAs 10A. As a result,
direct communication between a legacy STA 10B and a cognitive STA
10A become possible.
[0179] Simulations have shown that a channel utilization
enhancement of about 6 to 7 percent for existing IEEE 802.11
networks becomes possible due at least to the additional network
services offered by cognitive STAs 10A, while achieving high
channel throughput in non-ISM channels for cognitive STAs due to
better medium access coordination.
[0180] Described now are further enhancements to cognitive radio
techniques in accordance with additional embodiments.
[0181] In the final version of IEEE 802.11s amendment a so-called
common channel framework (CCF) is to be considered as a
non-compulsory technique to offer higher aggregate channel
throughput due to simultaneous multiple channel deployment and
concurrent data transmissions. In this approach mesh points (MPs)
are allowed to utilize a common channel to negotiate about an
available data channel that will be deployed for exchanging data
frame(s) between the source and the destination entities
(abbreviated by SE and DE). The negotiation phase is accomplished
by exchanging two designated control frames referred to as
ready-to-switch (RTX) and clear-to-switch (CTX).
[0182] It is noted in this regard that according to the IEEE
802.11s standard, any device that supports pre-defined mesh
services in a wireless mesh network (WMN) is called a mesh point
(MP). Note that a MP can be either a dedicated infrastructure
device or a user appliance that is able to fully participate in
both the formation and operation of the mesh network
simultaneously.
[0183] Further in this regard the source entity (SE) refers to any
type of mesh equipment that has something to send. This can be
either a MP or a mesh access point (MAP). In addition, a cognitive
source entity (CSE) may be considered to be a frequency agile SE.
The destination entity (DE) refers to any type of mesh equipment
that is intended to receive data frame(s) from a SE. Similarly, by
cognitive destination entity (CDE) what is intended is a frequency
agile DE.
[0184] At the beginning of the exchange, the SE sends an RTX frame
over the common channel to inform the DE of an intended data
transmission targeted to it. The SE offers an empty channel to the
DE which is to be deployed during the data transmission. Using a
dedicated field in the header of the RTX, i.e., "destination
channel information", the SE advertises the channel to the DE. When
the DE is also interested in deployment of the offered channel, it
responds using a CTX control frame which is also transmitted over
the shared common channel. Subsequent to the correct reception of
the CTX frame by the SE, both involved entities switch to the
destination channel in order to commence the agreed upon data
transmission. After switching to the destination channel, and after
a time period equal to DIFS (Distributed Inter-Frame Space), if the
media is sensed idle the SE will be allowed to start transmission
of data frame(s) to the DE. At the end, an acknowledgment (ACK)
frame is delivered through the destination channel back to the SE.
It is noted that there is a special type of MP, the mesh access
point (MAP), which serves also as an access point (AP) in addition
to providing pre-defined conventional mesh services.
[0185] In IEEE 802.11s it is assumed that the MPs are equipped with
a single radio transceiver. As a result, a particular MP on the
common channel is unable to communicate with other MPs which are
operating on the other channels. At the same time, single-radio MPs
on other channels are unaware of the network status on the common
channel.
[0186] As was discussed above, the cognitive IEEE 802.11 STAs 10A
are able to establish the cognitive basic service set (CBSS) to
which both legacy and cognitive STAs are able to join. All CBSSs
are established on the ISM frequency bands in which the legacy STAs
10B operate, enabling all legacy STAs 10B to detect, probe and
associate to the existing CBSSs. In addition, based on the above
concepts the cognitive STAs 10A are not allowed to deploy the
shared ISM channel for private data transmissions, except in the
case where a legacy STA 10B wants to establish a link layer
connection with a cognitive STA 10A, or when the destination of
interest of a cognitive STA 10A is a legacy STA 10B. In these two
cases the cognitive STA 10A operates on the shared ISM channel in
order to be able to exchange the intended data frame(s) with the
legacy STA 10B. By establishing the CBSS over a conventional ISM
channel, the cognitive STAs 10A are able to provide additional
network connectivity and packet forwarding to the legacy system on
one hand, and on the other hand they are allowed to utilize the
shared ISM channel as a common control channel that is primarily
exploited for cognitive-management/cognitive-control traffic
transmissions. Based on the above-referenced cognitive medium
access control protocol one may conclude that there is no need for
a common control channel to be particularly specified in order to
enable the cognitive STAs 10A to negotiate about the use of non-ISM
data channels.
[0187] The exemplary embodiments extend the foregoing MAC entity
protocol and operation to the IEEE 802.11s type of WMNs. In this
case a cognitive wireless mesh network (CWMN) may be defined as a
set of mesh entities (MEs) including both cognitive MPs/MAPs and
conventional or legacy 802.11s MPs/MAPs. In the CWMN ISM
band(s)/channel(s) are particularly utilized for 802.11s MPs/MAPs
and legacy 802.11 STAs traffic delivery, and for exchanging
cognitive-management and control frames to negotiate concerning
non-ISM channels that may be deployed. In this CWMN architecture
the ISM band(s)/channel(s) are deployed as common control channel
by all cognitive STAs, cognitive MPs (CMPs), and cognitive MAPs
(CMAPs). As a result, cognitive mesh entities (CMEs) 40 should
preferably avoid using ISM bands/channels for private data
transmissions as much as possible. On the other hand, and based on
the above architecture, non-ISM channels are specially utilized for
data transmissions in a distributed manner coordinated by cognitive
STAs, CMPs, and CMAPs.
[0188] As employed herein an ME encompasses any type of mesh
equipment that belongs to the WMN. This can be a simple IEEE 802.11
STA which has been already associated to the WMN, an MP, or an MAP.
In addition, an CME is any type of ME that exhibits frequency
agility, and may be a cognitive 802.11 STA, an CMP, or an CMAP.
[0189] FIG. 17 shows a simplified block diagram of a two
transceiver CME 40. Similar to the cognitive STA 10A having the
cognitive MAC entity protocol discussed above, each CME 40 is
equipped with a plurality of transceivers, namely an ISM
transceiver 42 and a non-ISM (cognitive) transceiver 44. The ISM
transceiver 42 is particularly dedicated to operate on the shared
ISM channel(s), while the non-ISM transceiver 44 functions as a
cognitive transceiver capable of switching between different
non-ISM frequency bands to handle data transmissions in a parallel
fashion. One result of this architecture is that the CME 40 does
not suffer from the problem of not being aware of the network
status. Note again in this regard that in IEEE 802.11s it is
assumed that the MPs are equipped with a single radio transceiver.
As a result an MP operating on the common channel is unable to
communicate with other MPs which are operating on the other
channels. At the same time, a single radio (single transceiver) MP
operating on one of the other channels is unaware of the network
status on the common channel.
[0190] Note that the shared ISM channel is especially dedicated to
only cognitive control and management purposes, and not to data
transmissions conducted by cognitive STAs/CMPs/CMAPs, i.e., CMEs.
It is assumed that there should be always an available common
control channel that is substantially free of any primary user
deployment, so that the cognitive radios are always allowed and
able to utilize it to negotiate about any possible deployable
non-ISM data channel for their private data transactions. Existing
ISM channels may be utilized by either independent neighboring
wireless LAN hotspots, IEEE 802.11 STAs associated with the WMN, or
legacy (i.e., non-cognitive) MPs/MAPs. The result is the deployment
of a CWMN that comprises both cognitive and non-cognitive MPs/MAPs
and various types of legacy IEEE 802.11 entities. The shared ISM
channel(s) are set aside particularly for use by legacy
non-cognitive 802.11/802.11s equipment.
[0191] Further in accordance with the exemplary embodiment, the
cognitive MAC entity 26 discussed above is incorporated into the
CME 40 as the MAC entity 46 (see FIG. 17).
[0192] FIG. 17, discussed above, shows a non-limiting example of a
CME 40. For the purposes of describing the exemplary embodiments, a
CDE and a CSE may each be constructed along the lines of the CME
40, as well as along the lines of the cognitive STA 10A shown in
FIG. 1B, and preferably include the MAC entity 26 that functions as
described above. Note that at one particular time a particular CME
40 may function as a CDE, and at another time it may function as a
CSE.
[0193] Of particular interest is the non-compulsory CCF which is
expected to be included in the final version of 802.11s amendment.
FIG. 18 illustrates the basic concept of CCF.
[0194] Regarding the IEEE 802.11s CCF scheme, and as was noted
above, both RTX and CTX control frames are transmitted over the
common channel. The main reason for exchanging these two control
frames on the common channel is that mesh entities, e.g. MPs/MAPs,
do not have any a priori knowledge about deployed and/or preferred
data channels of other mesh entities. As a result, source entities
are required to transmit RTX frames over the common channel, and
the destination entities of interest are expected to be waiting for
a link layer connection establishment request. The current IEEE
802.11s CCF scheme suffers at least from a lack of knowledge about
the network status when a tagged ME is currently operating on a
channel other than the common channel. As a result, there is no
guarantee that a particular RTX will lead to a successful link
layer connection establishment.
[0195] The exemplary embodiments not only apply the enhanced
cognitive MAC entity to the WMN, but also reduce the incurred
control overhead due to the channel negotiation phase. Recalling
that in IEEE 802.11s there is a particular common channel and a few
parallel/orthogonal data channels that are chiefly dedicated to
data transmission, the exemplary embodiments employ the ISM
channel(s)/band(s) as common channel(s), while all available
non-ISM frequency opportunities can be utilized for data
transmissions by the CMEs 40.
[0196] An important aspect of this feature is the way in which CMEs
40 are configured to operate on non-ISM channels in an efficient
fashion. In contrast to the cognitive 802.11 entities that were
discussed above, such as the cognitive STAs 10A, it is more
efficient for a CME 40 to select an available non-ISM frequency
opportunity, based on a set of pre-defined spectrum sensing
criteria, as its long-term residency channel (LTRC), and then
subsequently inform its one-hop neighbor MEs of the selected LTRC.
After choosing a particular non-ISM channel as the LTRC, the CME 40
keeps its non-ISM transceiver 44 tuned to the selected LTRC until,
based on some criterion or criteria, it becomes evident that
switching to another channel may be more beneficial. In the case of
permanent channel switching, the CME 40 is mandated to announce the
new non-ISM channel that is going to be utilized just after the
switching announcement. Basically, permanent channel switching is
announced using a designated frame referred to as a channel
switching (CHSW) frame.
[0197] When the CME 40 wishes to commence a data transmission to
one of its one-hop cognitive mesh neighbors (CMNs), due to the fact
that it already knows the LTRC of the intended DE, the negotiation
phase can be simply removed if the SE is also interested in the
current residency channel of the DE. To accomplish the intended
data transmission, the CME 40 switches its cognitive transceiver 44
from its LTRC to the LTRC of the DE. When the CME 40 decides to
switch its cognitive transceiver 44 to another non-ISM channel to
conduct a data transmission, it preferably reports this fact, and
the corresponding absence period from its LTRC.
[0198] When the CME 40 decides to switch to another channel for a
planned data transmission, it may use an eRTX frame sent over the
shared ISM channel. As employed herein the CMEs 40, instead of
using RTX/CTX, exchange new eRTX/eCTX control frames. The new
eRTX/eCTX frames are similar to conventional RTX/CTX frames, but
contain additional fields. Sending the eRTX frame over the shared
ISM channel simultaneously achieves two goals: 1) establishment of
a link layer connection with the intended DE, and 2) informing the
one-hop neighbors about the absence (on-leave) status (absence from
the LTRC) and its corresponding time duration. Note that the
overhead due to CMEs 40 is reduced by the use of only one control
frame (the eRTX control frame) to establish the intended link layer
connection and to report the on-leave status for a particular time
period.
[0199] By the use of these techniques MEs are provided with
frequency agility to obtain higher channel capacity and to
coordinate channel activities in such a way as to incur lowest
level of overhead in the common control channel. In addition,
on-leave information can be simply employed for destination non-ISM
channel (the LTRC of the DE) reservation purposes, in a manner
similar to the Duration/ID which is chiefly used for updating the
network allocation vector (NAV) in 802.11 ISM channels. Here, the
legacy IEEE 802.11 (and its 802.11s amendment) Duration/ID field of
the eRTX and eCTX control frames are used for the shared ISM
channel reservation (i.e., NAV update in ISM channel). In contrast,
the on-leave information carried by a designated field in eRTX/eCTX
is used for the destination non-ISM channel reservation in which
the targeted cognitive data transmission is intended to be
accomplished.
[0200] When the a priori knowledge about the CDE current LTRC is
correct, upon reception of an eRTX which carries the correct CDE
LTRC channel information the CDE can simply respond with an eCTX
frame. This does not occur over the shared ISM channel, but instead
on the current non-ISM LTRC of the CDE. In addition to eCTX, both
DATA and ACK frames are preferably also sent on the CDE's LTRC.
Based on this approach the usage of the shared ISM channel is
minimized, as it need be utilized only for cognitive control frame
exchange.
[0201] On the other hand, when the CSE has incorrect a priori
knowledge about the LTRC of the desired CDE, upon reception of the
eRTX frame by the CDE it becomes aware that the knowledge of the
CSE about its current LTRC is incorrect and, as a consequence, the
expected eCTX frame is delivered over the ISM channel. Afterwards
the CSE receives the eCTX from the shared ISM channel and is
informed of its incorrect knowledge concerning the LTRC of the
intended CDE. In this circumstance the CDE is expected to place the
correct information in the transmitted eCTX and, in a subsequent
step, the CSE sends another eRTX frame over the shared ISM channel,
but with the updated (and now correct) information. It can be noted
that reception of an eRTX frame, which carries incorrect
information, by entities that have no prior background knowledge of
the LTRC of the CDE distributes incorrect information through the
network. Thus, such undesirable cases are preferably addressed as
quickly as possible. By sending another eRTX frame the CSE prevents
distribution of inconsistent information through the wireless
system. The CSE is then allowed to commence its data transmission
over the LTRC of the CDE after sending the second eRTX, and an
additional SIFS.
[0202] Another issue that should be addressed properly is the case
where a CME decides to perform a temporary channel switching
operation while at least one of its neighbors has already initiated
backoff cycles intended for data transmission to the CME. Recall
that the switching entities are required to announce their
transition to the other channels in the form of eRTX frame if the
transition is intended for a data transmission. Upon reception of a
switching notification in the form of an eRTX, the cognitive entity
or entities that are conducting the backoff cycles preferably
suspend the counting down of backoff timers until the end of the
on-leave period. In the case of a permanent switching operation
there is no need to suspend the ongoing backoff cycle, since based
on the cognitive scheme, CMEs are enforced to follow two rules when
counting down their backoff counters:
[0203] 1. If the backoff counter is loaded by an integer value B,
for any subsequent countdowns before reaching `1` CME shall perform
carrier sensing only on the shared ISM channel;
[0204] 2. For the last countdown, i.e., when counting down from one
to zero, both the shared ISM and the destination non-ISM channels
are simultaneously sensed as being idle for at least a time
duration equal to DIFS (i.e., Distributed Inter-Frame Space).
[0205] An important difference between the conventional backoff
algorithm employed by IEEE 802.11s and the one described herein
leads to a lower access delay in comparison to the existing
multi-channel MAC protocols. In addition, the backoff technique is
more robust to the hidden terminal problem, as compared to the
802.11s common channel framework.
[0206] It can be noted that a WMN may be considered to be
substantially fixed in nature (i.e., experiencing little or no
topology alteration) and, as a result, the use of distributed radio
resource allocation techniques can be more effective in these kinds
of wireless systems as compared to infrastructure-less mobile ad
hoc networks. Although the mesh entities do not exhibit mobility
(or at least they are intended to have no mobility during a
long-term deployment), the wireless clients associated with the WMN
are totally free to roam between different 802.11 hotspots and
regional basic service areas (BSAs). Since 802.11s-based mesh
entities are unaware of the network status when they are operating
on a channel other than the common channel, both the multi-hop
nature of the wireless system and the client mobility can lead to
system throughput degradation and higher medium access delay. The
exemplary embodiments described below address these various
problems in a distributed/frequency-aware manner.
[0207] Described now are embodiments of the protocol core algorithm
and the corresponding frame structures. Based on the enhanced MAC
entity 26 and related features that were described above, the
channel information (CI) field is added to the header of control
802.11 frames (e.g., announcement traffic indication message
(ATIM), ready-to-send (RTS), clear-to-send (CTS), etc.) and
management 802.11 frames (e.g., Beacon, Probe Response, etc.) and
is used for advertising possibly deployable non-ISM channels. The
definition of the position of new header field is important so that
legacy 802.11 equipment are enabled to recognize/compile all
legacy/known header fields to enable, for example, their NAVs to be
correctly updated. In other words, the legacy STAs 10B should be
able to deduce all required information from the legacy fields, and
the new fields should be placed in such a way that they can be
simply discarded by the legacy 802.11 STAs. One preferred location
for the CI field is before the frame check sequence (FCS) field in
all control/management frames.
[0208] As was noted above, in IEEE 802.11s two designated control
frames, i.e., RTX and CTX, are used by the CCF scheme especially
for link layer connection establishment between neighboring MEs. At
least one aspect of the disclosed embodiments includes expanding
the RTX/CTX frame functionality to enable the CWMN to coordinate
cognitive concurrent data transmissions in an efficient manner.
These control frames are renamed as eRTX and eCTX, respectively
(see FIGS. 20A and 20B). Note that by introducing eRTX/eCTX it is
not intended to define any new frame type to the existing IEEE
802.11s standard. Instead the same frame types as for the legacy
RTX/CTX are used, but with additional fields. When the CME 40
desires to set up a link layer connection with another CME using
the cognitive common channel framework (CCCF), it sends out an eRTX
frame over the shared ISM channel in which the CWMN is already
established. There is no need for the CME 40 to advertise any
non-ISM channels to the intended CDEs when requesting a link layer
connection set up. In fact, each CME 40, based on a set of
pre-defined channel sensing criteria, chooses a non-ISM channel as
its LTRC and then tunes its non-ISM cognitive transceiver 44 to the
selected LTRC. For establishing a link layer connection with a CME,
the CSE switches to the LTRC of the CDE in order to accomplish the
data transmission. For this reason the CSE places the channel
information of the LTRC of the CDE within the eRTX to be delivered
on the shared ISM channel. The destination channel information is
included in a designated field, which may be referred to as
achannel switching information element (CHSWIE). All one-hop
neighbors of the CSE, by reception of the transmitted eRTX, are
informed of the time period during which the cognitive neighbor
will be on-leave. In fact, the CSE may be mandated to report its
on-leave duration since if it is assumed that every CME 40 is
equipped with only one non-ISM transceiver 44 that can be tuned to
only one channel at a given time. By transitioning to another
channel the CME 40 is unable to receive any data frames on its LTRC
and, as a result, all CMEs that are interested in transmitting data
frames to the channel switching CME need to be informed about the
situation. For example, a CME 40 that has already initiated a
backoff cycle for data transmission to another CME, which is going
to switch temporarily from its LTRC to a different channel, should
be informed to suspend the ongoing backoff cycle for the duration
of the on-leave period. In addition, and as was noted above, the
on-leave time duration is also employed for medium reservation and
NAV update in the destination channel to which the CME desires to
switch.
[0209] In accordance with this aspect, there are defined two
different types of channel switching that the CME 40 is allowed to
perform. A first type of channel switching is referred to as
temporary switching, and is accomplished when the CME 40 wants to
perform a data transmission on a non-ISM channel other than its
LTRC. The second type of channel switching is referred to as
permanent switching, and is conducted by the CME 40 based on a set
of pre-defined channel sensing criteria (e.g., it discovers that
switching to another non-ISM channel can be more beneficial). Due
to the fact that temporary channel switching is always initiated by
the CSE 40 when it desires to perform a data transmission, the eRTX
control frame is used to establish the intended link layer
connection and to also simultaneously inform one-hop CMEs of the
channel switching. Furthermore, the newly defined channel switching
(CHSW) control frame is used to report permanent switching.
[0210] FIG. 19 illustrates a simple frame flow to announce a
permanent switching of the CME 40 from channel C1 to channel C2.
The permanent channel switching announcement is made using the CHSW
transmitted on the shared ISM channel. As the first step, the
switching CME 40 initiates a backoff cycle and performs continuous
carrier sensing on the shared ISM channel. When the backoff counter
reaches zero, the CME 40 sends out a CHSW frame on the shared ISM
channel carrying the information of the non-ISM channel to which
the switching is to be accomplished. Since there is no particular
destination/receptor entity for the CHSW frame, the receiver
address (RA) in the CHSW may be loaded with, for example, the frame
initiator MAC address.
[0211] In addition to the aforementioned channel switching
approaches, which are basically initiated by the cognitive entity
(CE), another type of permanent channel switching is one initiated
by the CME 40 that desires to transmit a set of data frames to more
than one CDE. In this approach the CSE invites all of the intended
CDEs to congregate in a particular non-ISM channel to receive the
data frames in an integrated manner. In this way the CSE achieves a
higher transmission capacity since it is able to accomplish all
intended data deliveries using frame bursting. To attain this goal
the CSE sends out a designated frame, which may be referred to as a
switching invitation (SWinv) frame, on the shared ISM channel. The
channel capacity improvement based on this technique may be
referred to as welfare enhancement (WE).
[0212] It can be noted that this approach may also be employed for
link layer multicasting. For the case of multicasting, the
multicast MAC address is simply placed in the dedicated RA of the
SWinv control frame. However, for the case of WE with multiple
unicast data transmissions, multiple unicast MAC addresses are used
to invite individual CDEs 40 to switch to the channel of interest.
In the CCCF it is possible to combine unicast and multicast channel
switching invitations in a single SWinv control frame, as explained
below.
[0213] FIGS. 20A and 20B illustrate the frame structure of the eRTX
and eCTX, respectively. Similar to the legacy RTX and CTX control
frames defined in IEEE 802.11s, an eRTX/eCTX includes an initial 2
byte frame control and 2 byte Duration/ID fields. With regard to
how the Duration/ID field should be set for eRTX and eCTX, in
essence the Duration/ID field of eRTX and eCTX frames is specially
used for the shared ISM channel reservation and NAV update. On the
other hand, the on-leave duration information carried by eRTX/eCTX
frames, which reflects the on-leave status time duration of the
CSE, is used for destination non-ISM channel reservation
purposes.
[0214] The above-described new CHSWIE field is inserted into both
the eRTX and eCTX control frames just before FCS field. For eRTX
and eCTX frames the CHSWIE field has a variable length depending on
the application scenario. In addition, it should be noted that RA
in the eRTX frame can be loaded with either a unicast, a multicast,
or a broadcast MAC address, whereas the same field in the eCTX
frame can be loaded only with the unicast MAC addresses.
[0215] FIG. 21 shows the frame format of CHSW control frame. The
CHSW frame is used whenever the CME 40 decides to change its
preferred LTRC channel permanently. The switching CME 40 sends out
the CHSW control frame over the shared ISM channel, with RA field
loaded with the MAC address of the CME sending the CHSW control
frame. This strategy is somewhat similar to the `CTS to self`
technique defined in the IEEE 802.11g amendment, but with a
different application.
[0216] FIG. 22 illustrates the detailed structure of the CHSWIE
field used in eRTX, eCTX, and CHSW frames. Up to two CI sub-fields
are included in a single CHSWIE. The CI structure was described
above (see FIGS. 3 and 4 and the corresponding description of
same). The first byte of the CHSWIE is dedicated to channel control
(CC) in order to control the basic structure of the CHSWIE and its
contents. In the CC field the first several bits are assigned to
convey "CHSWIE status". Basically, if the CHSWIE has no appended CI
sub-field, then the CHSWIE status is set to 00 (i.e., Empty
CHSWIE); otherwise, the aforementioned bits are loaded with 11
(i.e., non-Empty CHSWIE). The next two successive bits in CC (i.e.,
the Proposition/Evaluation and Decision Bits) are not currently
used in CCCF and, together with the subsequent reserved bit, may be
used for future protocol development purposes. Note that the bits
of the Reason Code in CCCF have different meanings than those shown
in FIG. 22. The first bit, i.e. "Permanent/Temporary Switching
Flag", is used to specify whether the intended channel switching is
permanent or temporary. The second bit, i.e., "No. of Channel
Fields", specifies the number of CIs that are appended to the
CHSWIE. The last bit, i.e., "Application Bit", may be used in
conjunction with the preceding bits.
[0217] When there is only one CI within the CHSWIE, as shown in
FIG. 22, it may be used for permanent switching. In this case the
CI sub-field carries the information of the non-ISM channel to
which the permanent switching by the CME 40 is to occur. Generally,
the Regulatory Class, Channel Mask and the channel Center Frequency
are the three sub-fields that are included in every CI regardless
of application or type of switching. When the CME 40 decides to
permanently change its current LTRC and switch to another non-ISM
channel, it places the information descriptive of the new channel
within the CI of the CHSWIE appended in a CHSW frame, and sends it
over the shared ISM channel (see FIG. 19).
[0218] On the other hand when there are two CIs within the CHSWIE,
the CHSWIE field is intended to be used in either eRTX or eCTX
(i.e., temporary switching). In this case the first CI represents
the "other party non-ISM channel information" while the second CI
corresponds to the "local party non-ISM channel information". By
"other party" what is meant is the CME 40 to which the frame is
addressed, i.e., the frame receptor. By "local party" what is meant
is the frame initiator that is transmitting the frame over the
wireless channel. As an example, for an eRTX frame the "other
party" refers to the CDE and the "local party" refers to the CSE
that is delivering the eRTX. As another example, for an eCTX frame
the "other party" refers to the original CSE that desires to send
data frames to a destination of interest, while the "local party"
refers to the original CDE who is to receive data frames from the
CSE.
[0219] The CHSWIE is terminated by a two bytes representing the
on-leave duration (OLD) sub-field. The OLD sub-field specifies the
time duration of absence of the switching CME from its LTRC. For
temporary channel switching, the OLD sub-field is loaded with a
non-zero value between 0000 and FFFF (hexadecimal). In contrast,
when the CME is performing a permanent switch to another LTRC the
OLD sub-field is with FFFF. Basically, the Duration/ID field in
both the eRTX and eCTX is employed for the shared ISM channel
reservation and NAV update, while the OLD sub-field of the CHSWIE
in eRTX and eCTX is used not only for on-leave duration reporting,
but also for channel reservation and NAV update in the destination
channel (e.g., the LTRC of the CDE). Therefore, it is important to
define the way by which both the Duration/ID and OLD sub-fields are
tuned when exchanging eRTX/eCTX control frames. For this reason, a
consideration is made of all possible scenarios that can take place
when the CME 40 desires to commence a data transmission with
another CME based on different combinations of Reason Code bit
pattern and CHSWIE status. The Tables shown in FIGS. 23A and 23B
tabulate these diverse scenarios in an integrated fashion. Note in
this regard that the most significant bit (MSB) of the Reason Code
field specifies the type of switching, i.e., Temporary (TSW) or
Permanent (PSW). In the case of Temporary Switching, the second bit
determines the number of Channel Information (CI) fields located
between Channel Control (CC) and the On-Leave Duration (OLD) field.
Possible CI fields are other party and local party non-ISM CI. In
the case of Temporary Switching, when two CI fields are included
(i.e., when the second bit is `1`), the other party CI should be
always before the local party CI (i.e., frame initiator CI). For
example, for the case of an eCTX with two CI fields, the first CI
corresponds to the CSE (other party) and the second CI corresponds
to the CDE (local party or eCTX initiator) non-ISM channel
information, respectively. Note that in these Tables, in addition
to the content of existing CI fields (i.e., the carried
information), the reason for which the CI field is included is also
presented, where the content is underlined and the reason is placed
within brackets [ ]. Further, by "CME's channel information" is
meant "the CME's current non-ISM channel information".
[0220] For the case of the CHSWIE status=11 (FIG. 23A), when the
CSE has a priori knowledge of the LTRC of the CDE, either correct
or incorrect, its transmitted eRTX frame always carries a non-empty
CHSWIE. As a result, the CHSWIE status bits of the CC the in CHSWIE
are loaded with 11.
[0221] As a first exemplary case, assume that the Reason Code has
been set to 000. This refers to temporary switching with only one
appended CI sub-field. When this configuration is used in an eRTX
control frame, the CI carries LTRC channel information of the CDE.
To establish a link layer connection with a CME, the CSE needs to
switch to the LTRC of the intended CDE. By transmitting an eRTX
frame, the CSE sends its request to the CDE and, at the same time,
it informs its one-hop cognitive mesh neighbors of its temporary
transition to another non-ISM channel. In this case the LTRC
channel information of the CDE should be included in the CHSWIE of
the eRTX. On the other hand when the Reason Code is loaded with
010, not only the LTRC of the CDE but also the LTRC channel
information of the CSE is included in the CHSWIE of the eRTX, where
both appended LTRCs (CDE and CSE) are based on local knowledge of
the CSE. It should be noted that inclusion of the LTRC channel
information of the CDE in eRTX control frames is considered
mandatory, while enclosure of the LTRC channel information of the
CSE is considered to be optional. Also note that for the case of
000 the size of CHSWIE is seven bytes while for the case of 010 it
is eleven bytes.
[0222] For the case of an eCTX, when the Reason Code is loaded with
000, the single appended CI carries the LTRC of the CSE. Basically,
when the knowledge of the CDE concerning the LTRC channel
information of the CSE is incorrect, the technique to obtain the
correct information is to use integrate the LTRC channel
information of the CSE in the received eRTX. Actually, the correct
information can be obtained via the received eRTX if the CSE has
previously included its LTRC channel information in the preceding
eRTX. When the CDE notices that its knowledge about the LTRC of the
CSE is incorrect, it preferably makes the required corrections to
its local databases as soon as possible. In fact, the CDE will be
unable to be informed about its incorrect knowledge unless the CSE
includes its LTRC channel information in the eRTX frame to be sent
to CDE (recall that inclusion of the LTRC channel information of
the CSE in the eRTX is optional). In this way the CDE is enabled to
determine that its knowledge regarding the LTRC of the CSE is
incorrect and needs to be corrected. In addition to the importance
of local knowledge correction, the CDE is also mandated to inform
its one-hop CMNs of the correct LTRC channel information of the
CSE. In fact, it is possible that the CDE has already distributed
its incorrect knowledge among its one-hop CMNs, resulting in
further undesirable distribution of the inconsistent/incorrect
information. Therefore, when the CDE (by reception of an eRTX frame
carrying the LTRC channel information of the CSE) is notified of
its incorrect knowledge, it is required to place the correct
information within its eCTX and send it out over the shared ISM
channel.
[0223] As a next case, consider an eCTX with Reason Code equal to
001: temporary switching with only one appended CI sub-field. In
this case the CI carries LTRC channel information of the CDE. When
a CME transmits an eRTX frame over the shared ISM channel, it is
required to place the LTRC channel information of its intended CDE
within the CHSWIE field of the eRTX. If the CDE receives the
delivered eRTX and notices erroneous appended information
concerning its LTRC, instead of sending the eCTX on its non-ISM
LTRC it instead sends out the eCTX over the shared ISM channel. In
addition the CDE places its correct LTRC channel information in the
eCTX to inform the CSE of its incorrect knowledge. In response, the
CSE responds with another eRTX on the ISM channel that contains the
LTRC correct information. Only after transmission of an eRTX with
the correct channel information is the CSE allowed to commence its
data delivery over the intended non-ISM channel (i.e., the LTRC of
the CDE).
[0224] For the case of an eCTX carrying a Reason Code loaded with
010, not only is the knowledge of the CSE about the LTRC of the CDE
incorrect, but the knowledge of the CDE about the LTRC of the CSE
is also incorrect. In other words, this case includes both
above-described scenarios, i.e., eCTX/000 and eCTX/001. The CDE is
required to send the eCTX with both the LTRC correct channel
information for the CSE and its own LTRC correct channel
information. This frame is delivered on the shared ISM channel, and
the CSE is then also required to send another eRTX frame over the
shared ISM channel that contains the correct LTRC channel
information for the CDE. The CSE then begins sending data frame(s)
on the LTRC of the CDE after completion of sending the second eRTX
(plus an additional SIFS).
[0225] Discussed now is the case of CHSWIE status=00 (FIG.
23B).When the CSE has no a priori knowledge about the LTRC of the
CDE, either correct or incorrect, its transmitted eRTX frame
carries an empty CHSWIE. As a result, the CHSWIE status bits of the
CC in the CHSWIE are loaded with 00. In addition, and as far as the
receiver side (i.e., CDE) is concerned, when both the appended LTRC
channel information of the CDE in the eRTX, and the knowledge of
the CDE about the LTRC of the CSE are correct, the CDE is allowed
to respond by the use of an eCTX transmitted on its own non-ISM
LTRC. In this case the CHSWIE of the eCTX is empty, with no further
appended CI field.
[0226] FIG. 24 shows the regular frame exchange between CSE and CDE
when the knowledge of the CSE concerning the LTRC channel
information of the CDE is correct. The CSE first initiates a
backoff cycle and commences counting down the backoff counter.
Before reaching one, for every countdown CSE is required to perform
carrier sensing only on the shared ISM channel. When counting down
from one to zero the CSE is expected to perform carrier sensing on
both the ISM and on the LTRC non-ISM channels of the CDEs. Upon
reaching zero the CSE sends an eRTX over the shared ISM channel.
Since in this scenario the included information in the eRTX
regarding LTRC of the CDE is correct, the CDE responds with an eCTX
on its LTRC channel. Subsequently, data and ACK frames are also
exchanged over the LTRC of the CDE. As was mentioned above, the
scenario shown in FIG. 24 corresponds to the case where the
knowledge of the CSE about the LTRC of the CDE is correct. In this
scenario the CSE includes only the LTRC information of the CDE in
the CHSWIE field of the eRTX which is being transmitted over the
shared ISM channel. When only the LTRC channel information of the
CDE is included in the eRTX the Duration/ID field in the eRTX is
loaded by eCTX (with 7 bytes CHSWIE)+2.times.SIFS, while the OLD
sub-field in the CHSWIE is loaded by eCTX (with 1 byte
CHSWIE)+DATA+ACK+3.times.SIFS. In this scenario the transmitted
eCTX on LTRC of the CDE carries an empty CHSWIE for which the
Duration/ID field is loaded by 00, and the OLD sub-field in the
CHSWIE is loaded by DATA+ACK+2.times.SIFS. The reason that the CSE
reserves the ISM channel for a time period equal to eCTX (with 7
bytes CHSWIE)+2.times.SIFS is due to the fact that the CSE should
take into account the case when its knowledge of the LTRC of the
CDE is totally incorrect, and thus where the intended CDE responds
with an eCTX on the shared ISM channel accompanied by a non-empty 7
byte CHSWIE with the correct LTRC channel information for the CDE.
In this case the shared ISM channel should have been reserved by
CSE beforehand to prevent any possible loss of channel control. If
the ISM channel is reserved for less than the aforementioned
period, it is possible that another ME may acquire control of the
shared ISM channel, and the above tagged CSE then needs to
re-initiate the entire eRTX/eCTX negotiation phase from the
beginning.
[0227] FIG. 25 shows the case when only the LTRC channel
information of the CDE is included in the eRTX, while the knowledge
of the CSE of the LTRC of the CDE is incorrect.
[0228] In this case the CSE appends only the LTRC channel
information for the intended CDE in the CHSWIE of the eRTX. In
addition, the knowledge of the CSE about LTRC of the CDE is
incorrect. As a result, the CDE responds with an eCTX on the shared
ISM channel to inform the CSE of its incorrect knowledge regarding
the LTRC of the CDE. When the CSE receives the eCTX from the shared
ISM channel it is accordingly notified about its incorrect
knowledge and, as a result, it sends out another eRTX on the common
ISM channel with corrected LTRC information of the CDE. Upon
successful reception of the second eRTX with the corrected LTRC
information, and after an additional SIFS, the CSE is allowed to
start transmission of its data frames on the (correct) LTRC of the
CDE. As it was pointed out above, when only the LTRC channel
information of the CDE is included in eRTX the Duration/ID field in
the eRTX is loaded with the eCTX (with 7 bytes
CHSWIE)+2.times.SIFS, while the OLD sub-field in the CHSWIE is
loaded with eCTX (with 1 byte CHSWIE)+DATA+ACK+3.times.SIFS. On the
other hand, based on this scenario the transmitted eCTX on the
shared ISM channel carries a CHSWIE loaded with the correct LTRC
channel information of the CDE for which the Duration/ID field is
loaded with eRTX (with 7 bytes CHSWIE)+SIFS and the OLD sub-field
in the CHSWIE is loaded with eRTX (with 7 bytes
CHSWIE)+DATA+ACK+3.times.SIFS. For the second eRTX the Duration/ID
field is loaded with 00, while the OLD sub-field is loaded with
DATA+ACK+2.times.SIFS.
[0229] FIG. 26 shows the channel reservation using eRTX and eCTX
when the CSE has no a priori knowledge of the LTRC of the CDE. When
the CSE has no a priori knowledge about the LTRC of the intended
CDE it simply sends an eRTX frame with an empty CHSWIE on the
shared ISM channel to inform the CDE of its desire to establish a
link layer connection. Upon reception of an eRTX frame with an
empty CHSWIE the CDE replies with an eCTX on the shared ISM channel
carrying its current LTRC channel information. When the CSE
receives the eCTX it sends another eRTX on the shared ISM channel
accompanied by LTRC channel information of the CDE in the CHSWIE
field. After completion of second eRTX transmission the CSE
commences data delivery on the targeted non-ISM channel (i.e., on
the LTRC of the CDE).
[0230] In the case depicted by FIG. 26 the Duration/ID field of
first eRTX is loaded with eCTX (with 7 bytes CHSWIE)+2.times.SIFS
while the OLD sub-field in the CHSWIE is loaded with eCTX (with 7
bytes CHSWIE)+eRTX (with bytes CHSWIE)+DATA+ACK+4.times.SIFS. The
transmitted eCTX on the shared ISM channel carries a CHSWIE loaded
with the LTRC channel information of the CDE for which the
Duration/ID field is loaded with eRTX (with 7 bytes CHSWIE)+SIFS
and the OLD sub-field in the CHSWIE is loaded with eRTX (with 7
bytes CHSWIE)+DATA+ACK+3.times.SIFS. For the second eRTX the
Duration/ID field is simply loaded with 00, while the OLD sub-field
is loaded with DATA+ACK+2.times.SIFS.
[0231] FIG. 27 shows the case where the CSE includes both CDEs and
LTRC channel information in the eRTX frame to be delivered on the
shared ISM channel. In this case it is assumed that the appended
LTRC channel information of the CDE in the eRTX is correct, while
the knowledge of the CDE about the LTRC of the CSE is incorrect.
Since the CSE has included its LTRC channel information in the
eRTX, the CDE is informed that its knowledge concerning the LTRC of
the CSE is incorrect. As a result, the CDE responds using an eCTX
delivered on the shared ISM channel appended with the correct LTRC
channel information of the CSE. In response, and after a SIFS, the
CSE commences transmission of data frames over the LTRC of the
CDE.
[0232] For the case shown in FIG. 27 the Duration/ID field of eRTX
is loaded with the eCTX (with 11 bytes CHSWIE)+2.times.SIFS, while
the OLD sub-field in the CHSWIE is loaded with eCTX (with 1 byte
CHSWIE)+DATA+ACK+3.times.SIFS. The transmitted eCTX on the shared
ISM channel carries a CHSWIE loaded with the LTRC channel
information of the CSE, for which the Duration/ID field is loaded
with `00` while the OLD sub-field in the CHSWIE is loaded with
DATA+ACK+2.times.SIFS.
[0233] FIGS. 28 and 29 illustrate all possible CHSWIE
configurations for both eRTX and eCTX, including their Duration/ID
and OLD sub-field setups.
[0234] As was described above, a designated control frame referred
to as SWinv is defined to be employed for Unicast Welfare
Enhancement (UWE) and Multicast Welfare Enhancement (MWE). Note
also that the SWinv control frame may be used as well for inviting
cognitive members of a multicast group to gather or congregate in a
particular non-ISM channel (which may be totally different than
their own respective LTRC channels). Based in large part on the UWE
concept the CME 40 is allowed to invite its intended CDEs to gather
in a certain non-ISM channel. If the invitation of the CSE is
accepted by the entire group of CDEs the CME 40 (i.e., the CSE in
this case) is enabled to use frame bursting to achieve a
significantly higher channel throughput (channel utilization) by
successively transmitting data frames addressed to the invited
CDEs. For the multicasting case (i.e., MWE), and since the CME 40
is not allowed to deploy the shared non-ISM channel for cognitive
unicast/multicast data transmission, and further since there is no
way to perform multicasting over multiple non-ISM channels at the
same time (due to the presence of the single non-ISM transceiver
44), the multicast CSE invites members of its intended multicast
group to gather in the particular non-ISM channel.
[0235] FIG. 30 illustrates the detailed frame format of the SWinv
and its CHSW IE. It may be noticed that the design of the SWinv
frame structure enables legacy 802.11 and 802.11s equipment to
interpret all of the appended legacy fields without difficulty.
Note that between the Duration/ID and transmitter address (TA)
fields the first receiver address (RA) is positioned. This address
can be a unicast, a multicast, or a broadcast MAC address. In
CHSWIE field up to two CI fields may be included, depending on the
application and addressing scenario. Subsequent to the CHSWIE field
up to three further RAs may be included. These RAs can be either
unicast or multicast MAC addresses. The SWinv is terminated with a
conventional FCS. TA represents the MAC address of the inviting CME
40, while the RAs hold invited CME MAC addresses. In FIG. 30 the CC
sub-field of CHSWIE in the SWinv is also illustrated. The first two
bits of the CC are dedicated to a "No. of Extra RA Fields". Using
these two bits the number of additional RA fields that are included
in the SWinv after the CHSWIE is specified to the receiver. For
example, 00: No extra RA fields, 01: one extra RA fields, 10: two
extra RA fields, 11: three extra RA fields. Several following bits
are presently not used. As before, the Permanent/Temporary
Switching Flag, No. of Channel Fields, and Application Bit are used
to differentiate between diverse unicast/multicast scenarios. The
Permanent/Temporary Switching Flag specifies the type of switching
for the case of multicasting. It should be noted that for unicast
switching (i.e., UWE), channel switching is always performed
permanently. As a result the OLD sub-field in the CHSWIE is not
needed to declare the amount of time invited CDEs are required to
switch to the destination non-ISM channel. Therefore, if the
planned switching is intended for only the UWE, the OLD sub-field
may be simply loaded by FFFF Hex. On the other hand, for
multicast-related switching scenarios the OLD sub-field may be set
to any needed value. If the Permanent/Temporary Switching Flag is
loaded with a one, then the OLD sub-field is preferably loaded with
FFFF Hex.
[0236] Note that by the use of the SWinv it is possible to combine
multicast and unicast switching invitations into a single control
frame. As a result the incurred overhead due to successive
switching invitations can be reduced. Since it is possible to
append both multicast and unicast switching invitations into a
single SWinv at the same time, a plurality of CI fields are used to
carry the destination channel related information for both the
multicast and unicast cases.
[0237] A description is now provided of exemplary different use
case applications of the SWinv control frame.
[0238] A first use case relates to multicast temporary channel
switching. When the CME 40 intends to invite members of a multicast
group to switch temporarily to a particular non-ISM channel it
sends a SW inv with No. of Extra RA Fields set to `00`,
Permanent/Temporary Switching Flag 15 set to `0` (indicating
Temporary), No. of Channel Fields 16 set to `0`, and the OLD
sub-field loaded with a value between 0 and FFFF Hex. The multicast
physical (MAC) address is placed in the first RA field (the RA
field between the Duration/ID and TA fields). The destination
non-ISM channel information is loaded into the single appended CI
sub-field in CHSWIE. OLD is loaded by MULTICAST_DATA+SIFS. The
transmitter address (TA) is loaded with the CSE (inviting ME) MAC
address. The Duration/ID field of SWinv is loaded with `00`. Upon
transmission of the SWinv on the shared ISM channel, plus an extra
SIFS, the multicast CSE commences delivering the multicast data
frame(s) on the destination non-ISM channel. No acknowledgment is
required to be returned by CDEs after completion of multicast data
transmission. In addition, invited CDEs are not required to
transmit a CHSW control frame when switching to the intended
destination channel. FIG. 31 shows the timing and frame exchange
pattern for multicast invitation/transmission in more detail.
[0239] In FIG. 31 the multicast CSE commences a backoff cycle based
on the legacy DCF access scheme. Before the backoff counter reaches
one the CSE performs carrier sensing only on the shared ISM
channel, while when counting down from one to zero it performs the
carrier sensing procedure for both the shared ISM and the
destination non-ISM channels. When the channels are sensed as being
idle for a period of time equal to at least DIFS, the CSE transmits
a SWinv on the shared ISM channel to invite two members of a
multicast group (i.e., D1 and D2) to switch to the destination
non-ISM channel, C3. Both invited CDEs switch to the C3 channel to
receive the multicast data frame. Upon completion of data
reception, both CDEs switch back to their initial LTRCs.
[0240] A second use case relates to multicast permanent channel
switching. When the CME 40 intends to invite members of a multicast
group to switch permanently to a particular non-ISM channel, it
sends a SWinv with No. of Extra RA Fields set to `00`,
Permanent/Temporary Switching Flag set to >1=, and No. of
Channel Fields set to `0`. The multicast MAC physical address is
placed in the first RA field. The destination non-ISM channel
information is loaded into the single appended CI sub-field in the
CHSW IE. The TA is loaded with the CSE (inviting ME) MAC address.
No extra RA field is appended after the CHSW IE. In this case, two
different strategies may be used for multicast permanent channel
switching.
[0241] In the first strategy, which may be referred to as Mode I
Multicast Permanent Channel Switching, the multicast CSE initiates
a backoff cycle based on the legacy DCF access scheme. Before the
backoff counter reaches one the CSE performs carrier sensing only
on the shared ISM channel, while when counting down from one to
zero the CSE performs the carrier sensing procedure for both the
shared ISM and the destination non-ISM channels. When the channels
are sensed idle for at least DIFS, the CSE transmits a SWinv on the
shared ISM channel to invite multicast members to switch
(permanently) to the destination non-ISM channel. In this mode the
Duration/ID field is loaded with 00. In addition, the OLD sub-field
of the CHSWIE in the SWinv is also loaded with 00. To inform the
CDEs that the intended multicast permanent channel switching is
initiated based on Mode I the CSE loads the Application Bit of the
CC sub-field in the SWinv control frame with zero. Upon completion
of the SWinv transmission, the CSE loads a switching timeout timer
with a maximum possible busy period in the shared ISM channel (see
IEEE 802.11/1999 standard) and waits for the CDEs to switch to the
destination channel. Whenever an invited CDE decides to switch to
the destination channel, it is required to perform carrier sensing
on the shared ISM channel. Each invited CDE sends a CHSW frame over
the ISM channel to inform its one-hop cognitive mesh neighbors (and
specially the multicast CSE) that it is permanently switching to
the new channel, and is thus selecting it as its new LTRC. Upon
expiration of the switching timeout timer CSE commences
transmission of multicast data frames using a regular RTX control
frame with the RA loaded with the multicast MAC address. For
transmission of the RTX frame the CSE is performs carrier sensing
on both shared ISM and the destination non-ISM channels. FIG. 32
illustrates the above described interactions in more detail.
[0242] In the second strategy, which may be referred to as Mode II
Multicast Permanent Channel Switching (Fast CHSW), the multicast
CSE initiates a backoff cycle based on the legacy DCF access
scheme. Before the backoff counter reaches one the CSE performs
carrier sensing only on the shared ISM channel, while when counting
down from one to zero the CSE performs the carrier sensing
procedure for both the shared ISM and the destination non-ISM
channels. When the channels are sensed idle for at least DIFS, the
CSE transmits a SWinv on the shared ISM channel to invite multicast
members to permanently switch to the destination channel. In this
second mode the Duration/ID field of the SWinv is loaded with 00
while the OLD sub-field of the CHSWIE is loaded with
MULTICAST_DATA+SIFS. To inform the CDEs that the intended multicast
permanent channel switching is initiated based on Mode II the CSE
loads the Application Bit of the CC sub-field in the SWinv control
frame with a one. In this mode, after completion of SWinv
transmission plus an SIFS, all CDEs are required to switch to the
destination non-ISM channel. Switching to the new non-ISM channel
is accomplished before informing one-hop cognitive mesh neighbors
of selecting the destination channel as the new LTRC using the
designated CHSW control frames. In parallel with the multicast data
frame reception the multicast CDEs may also contend for the shared
ISM channel to transmit the required CHSW frames to inform their
one-hop neighbors of the selection of the non-ISM channel as their
new LTRC. In this mode there is no need for the CSE to initiate the
switching timeout timer. This type of multicast switching is well
suited for use in fast channel switching cases, such as when the
multicast MSDU(s) are sensitive to the incurred access/transmission
delay (i.e., for voice or video traffic). FIG. 33 illustrates this
above-described use case.
[0243] A third use case relates to unicast welfare enhancement
(UWE). When the CME 40 intends to invite another single CME 40 to
permanently switch to a particular non-ISM channel it sends a SW
inv with the No. of Extra RA Fields set to 00, the
Permanent/Temporary Switching Flag set to 1, the No. of Channel
Fields set to 0 and the OLD sub-field loaded with FFFF Hex. When
the CME intends to invite a plurality of other CMEs 40 to
permanently switch to a particular non-ISM channel it sends a SWinv
with the No. of Extra RA Fields loaded with 01, 10, or 11, the
Permanent/Temporary Switching Flag set to 1, the No. of Channel
Fields set to 0 and the OLD sub-field loaded with FFFF Hex. The
destination non-ISM channel information is loaded into a single
appended CI sub-field in the CHSWIE. The TA is loaded with the
address of the CSE (inviting ME). The Duration/ID field of SWinv is
loaded with 00. By the use of the SWinv the CSE is able to invite
up to four CMEs to switch to the destination non-ISM channel. In
this case the Application Bit in the CC sub-field is not used.
Subsequent to receiving the SWinv the invited CDEs contend for the
shared ISM channel to transmit CHSW frames in order to inform their
one-hop cognitive neighbors of the permanent switch to the new
non-ISM channel.
[0244] A fourth use case relates to combined Multicast/Unicast
channel switching. In this case the CSE invites a multicast group
and a set of unicast CMEs to permanently switch to either one or
two non-ISM channel(s). The multicast MAC address is loaded into
the first RA field between Duration/ID and TA fields, while the
remaining three RA addresses may be used for the unicast channel
switching cases. In addition, the first CI sub-field in the CHSWIE
is used for multicast channel switching and the second CI is
utilized for the unicast channel switching case. As in the use
cases described above, the Duration/ID field is loaded with 00. The
OLD sub-field in the CHSWIE is loaded with either 00 or
MULTICAST_DATA+SIFS, depending on the type of multicasting channel
switching: Mode I or Mode II. Using the Application Bit in the CC,
the type (or mode) of the multicast switching is specified. Based
on the specified type of multicast switching the OLD sub-field is
loaded with the appropriate value, i.e., 00 (Mode I) or
MULTICAST_DATA+SIFS (Mode II). The subsequent channel activity is
the same as multicast switching explained above, and is coordinated
based on the switching mode: I or II. The OLD sub-field is not used
for the appended unicast case, instead it is used for multicast
switching and loaded according to the multicast mode. The unicast
CDEs are required to transmit a CHSW frame over the shared ISM
channel if they agree to switch their LTRC to the advertised
destination channel. FIG. 34 illustrates an example for this
case.
[0245] It can be appreciated that the embodiments described with
reference to FIGS. 18-34 provide a further novel frequency agile
medium access control protocol capable of coordination of
concurrent multi-channel data communications in a distributed
fashion. These embodiments may employ the distributed multi-channel
cognitive MAC protocol for the 802.11 wireless LANs (an enhanced
MAC or eMAC) which was described in reference to FIGS. 1-17. These
embodiments provide unicast/multicast welfare enhancement, and a
practical approach to reduce the incurred overhead on the shared
ISM channel due to cognitive mesh entities control/management frame
exchange. In addition, medium access delay experienced by cognitive
mesh entities is reduced.
[0246] Based on the foregoing it should be apparent that the
exemplary embodiments provide a method, apparatus and computer
program product(s) to enhance the operation of wireless networks
that include cognitive radio apparatus, such as mobile stations and
mesh elements.
[0247] FIG. 35 is a logic flow diagram that illustrates the
operation of a method, and a result of execution of computer
program instructions, in accordance with the exemplary embodiments.
At Block 35A there is a step of sending a message from a first
cognitive radio apparatus to at least one second cognitive radio
apparatus, the message being sent over a first communication
channel and comprises an advertisement of at least one second
communication channel for use in sending data from the first
cognitive radio apparatus to the at least one second cognitive
radio apparatus. At Block 35B there is a step of receiving a reply
from the at least one second cognitive radio apparatus over the
first communication channel, the reply comprising one of an
acceptance of one of the at least one second communication
channels, a rejection of the at least one second communication
channel and an advertisement of at least one third communication
channel, or a rejection of the at least one second communication
channel without an advertisement of at least one third
communication channel. At Block 35C there is a step of transmitting
the data from the first cognitive radio apparatus to the at least
one second cognitive radio apparatus over an agreed upon one of the
second or third channels.
[0248] In the method, and the result of execution of computer
program instructions as in the previous paragraph, where in a case
where the reply is a rejection of the at least one second
communication channel and the advertisement of at least one third
communication channel, further comprising sending a response from
the first cognitive radio apparatus to the second cognitive radio
apparatus over the first communication channel, the response
comprising one of an acceptance of one of the at least one third
communication channels or a rejection of the at least one third
communication channel.
[0249] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the reply
comprises one of a reason for the acceptance or a reason for the
rejection.
[0250] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the first
communication channel is a common channel in an ISM frequency band
that is used by cognitive radio apparatus and by non-cognitive
radio apparatus.
[0251] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the first
cognitive radio apparatus operates in one of a power saving mode
enabled state or a power saving mode disabled state.
[0252] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the first
cognitive radio apparatus comprises a first transceiver for
communication over the first communication channel and a second
frequency agile transceiver for communication over the second or
third communication channels, further comprising a first task list
associated with the first transceiver and a second task list
associated with the second transceiver, each task list comprising
at any given time one or both of reception tasks and transmission
tasks and, for at least the second task list, a communication
channel associated with each task.
[0253] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where tasks in
a given one of the task lists are prioritized, where reception
tasks are assigned a higher priority than transmission tasks.
[0254] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use, if possible, a same
communication channel for more than one task.
[0255] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use, if possible, a same
communication channel for two reception tasks.
[0256] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use, if possible, a same
communication channel for two transmission tasks.
[0257] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use, if possible, a same
communication channel for a reception task and for a transmission
task.
[0258] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use a communication channel
for a multicast transmission task to a plurality of second
cognitive radio apparatus.
[0259] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising scheduling tasks so as to use a communication channel
for a multicast transmission task to a plurality of second
cognitive radio apparatus, and the same communication channel for a
unicast transmission to at least one further cognitive radio
apparatus.
[0260] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the
second and third communication channels are in a non-ISM frequency
band, and where a decision to accept or reject a particular channel
is based at least in part on a result of spectrum sensing in the
non-ISM frequency band to detect an appearance of a primary
user.
[0261] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where an
advertised communication channel is accepted if it is not a channel
in which a primary user appeared, and it has satisfactory spectrum
quality results.
[0262] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising operating a cognitive medium access control entity that
comprises a Last Successfully Experienced Channel (LSEC) table, a
Primary User Appearance (PUA) table, and a Transient Zone that
buffers identifications of locally and remotely discovered non-ISM
channels, where each channel identified in LSEC and PUA table
entries includes an associated time index having a value that
changes periodically and that controls potential usage of the
channels identified in the LSEC and PUA tables.
[0263] In the method, and the result of execution of computer
program instructions as in the previous paragraph, where the value
of the time index is changed every Beacon Interval.
[0264] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the time
index value of a particular channel is reset to zero when the
channel is placed in the LSEC table, where the time index value is
periodically incremented, and where the channel is not available to
be reused until the time index value reaches some predetermined
non-zero value.
[0265] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the time
index value of a particular channel is set to some predetermined
value when the channel is placed in the PUA table, where the time
index value is periodically decremented, and where the channel is
removed from the PUA table when the time index value reaches
zero.
[0266] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where
advertisements of the second and third communication channels each
comprise fields for specifying an identification of a center
frequency and a regulatory class.
[0267] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the first
cognitive radio apparatus comprises a first transceiver for
communication over the first communication channel, where the first
communication channel is a common channel in an ISM frequency band
that is used by cognitive radio apparatus and by non-cognitive
radio apparatus, and a second frequency agile transceiver for
communication over at least one communication channel in a non-ISM
frequency band, further comprising sending a message over the first
communication channel to the at least one second cognitive radio
apparatus that also comprises first and second transceivers, the
message instructing the second cognitive radio apparatus to one of
switch temporarily or permanently to a particular communication
channel in the non-ISM frequency band.
[0268] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where in one
mode of operation the second cognitive radio apparatus acknowledges
the message over the first communication channel before switching
to the particular communication channel in the non-ISM frequency
band, and where in another mode of the operation the second
cognitive radio apparatus acknowledges the message over the first
communication channel after switching to the particular
communication channel in the non-ISM frequency band.
[0269] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, further
comprising sending at least one of multicast data and unicast data
over the particular communication channel in the non-ISM frequency
band.
[0270] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the
message invites a multicast group of second cognitive radio
apparatus to switch temporarily to a particular non-ISM
communication channel.
[0271] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the
message invites a multicast group of second cognitive radio
apparatus to switch permanently to a particular non-ISM
communication channel.
[0272] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the
message invites a single second cognitive radio apparatus to switch
temporarily to a particular non-ISM communication channel.
[0273] In the method, and the result of execution of computer
program instructions as in the previous paragraphs, where the
message invites a multicast group of second cognitive radio
apparatus and a single second cognitive radio apparatus to switch
permanently to one or more than one particular non-ISM
communication channels.
[0274] The various blocks shown in FIG. 35 may be viewed as method
steps, and/or as operations that result from operation of computer
program code, and/or as a plurality of coupled logic circuit
elements constructed to carry out the associated function(s).
[0275] In general, the various exemplary embodiments may be
implemented in hardware or special purpose circuits, software,
logic or any combination thereof. For example, some aspects may be
implemented in hardware, while other aspects may be implemented in
firmware or software which may be executed by a controller,
microprocessor or other computing device, although the disclosed
embodiments are not limited thereto. While various aspects of the
exemplary embodiments may be illustrated and described as block
diagrams, flow charts, or using some other pictorial
representation, it is well understood that these blocks, apparatus,
systems, techniques or methods described herein may be implemented
in, as non-limiting examples, hardware, software, firmware, special
purpose circuits or logic, general purpose hardware or controller
or other computing devices, or some combination thereof.
[0276] As such, it should be appreciated that at least some aspects
of the exemplary embodiments may be practiced in various components
such as integrated circuit chips and modules. The design of
integrated circuits is by and large a highly automated process.
Complex and powerful software tools are available for converting a
logic level design into a semiconductor circuit design ready to be
fabricated on a semiconductor substrate. Such software tools can
automatically route conductors and locate components on a
semiconductor substrate using well established rules of design, as
well as libraries of prestored design modules. Once the design for
a semiconductor circuit has been completed, the resultant design,
in a standardized electronic format (e.g., Opus, GDSII, or the
like) may be transmitted to a semiconductor fabrication facility
for fabrication as one or more integrated circuit devices.
[0277] Various modifications and adaptations to the foregoing
exemplary embodiments may become apparent to those skilled in the
relevant arts in view of the foregoing description, when read in
conjunction with the accompanying drawings. However, any and all
modifications will still fall within the scope of the non-limiting
and exemplary embodiments.
[0278] For example, while the exemplary embodiments have been
described above in the context of the IEEE 802.11 type of system,
it should be appreciated that the exemplary embodiments are not
limited for use with only this one particular type of wireless
communication system, and that they may be used to advantage in
other wireless communication systems. Further, all of the various
specific references to specific frequency bands and channels and
channel numbers, the number of bits in certain frame fields, the
names of these certain bits and fields, the ordering of these
fields, the number of certain fields within a given frame and the
like are meant to be exemplary, and are not to be construed as
limitations upon the implementation and practice of the various
exemplary embodiments.
[0279] It should be noted that the terms "connected," "coupled," or
any variant thereof, mean any connection or coupling, either direct
or indirect, between two or more elements, and may encompass the
presence of one or more intermediate elements between two elements
that are "connected" or "coupled" together. The coupling or
connection between the elements can be physical, logical, or a
combination thereof. As employed herein two elements may be
considered to be "connected" or "coupled" together by the use of
one or more wires, cables and/or printed electrical connections, as
well as by the use of electromagnetic energy, such as
electromagnetic energy having wavelengths in the radio frequency
region, the microwave region and the optical (both visible and
invisible) region, as several non-limiting and non-exhaustive
examples.
[0280] Furthermore, some of the features of the various
non-limiting and exemplary embodiments may be used to advantage
without the corresponding use of other features. As such, the
foregoing description should be considered as merely illustrative
of the principles, teachings and exemplary embodiments, and not in
limitation thereof.
* * * * *