U.S. patent application number 16/167034 was filed with the patent office on 2019-06-27 for network-aware adjacent channel interference rejection and out of band emission suppression.
This patent application is currently assigned to University of South Florida. The applicant listed for this patent is Huseyin Arslan, Selcuk Kose, Berker Pekoz. Invention is credited to Huseyin Arslan, Selcuk Kose, Berker Pekoz.
Application Number | 20190199383 16/167034 |
Document ID | / |
Family ID | 66949669 |
Filed Date | 2019-06-27 |
![](/patent/app/20190199383/US20190199383A1-20190627-D00000.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00001.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00002.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00003.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00004.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00005.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00006.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00007.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00008.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00009.png)
![](/patent/app/20190199383/US20190199383A1-20190627-D00010.png)
View All Diagrams
United States Patent
Application |
20190199383 |
Kind Code |
A1 |
Pekoz; Berker ; et
al. |
June 27, 2019 |
NETWORK-AWARE ADJACENT CHANNEL INTERFERENCE REJECTION AND OUT OF
BAND EMISSION SUPPRESSION
Abstract
A system and method for adaptively utilizing transmitter
windowing, receiver windowing and alignment signals for minimizing
interference and maximizing capacity and energy efficiency based
upon the received power ratios of links in adjacent bands of a
cellular communication network.
Inventors: |
Pekoz; Berker; (Tampa,
FL) ; Kose; Selcuk; (Tampa, FL) ; Arslan;
Huseyin; (Tampa, FL) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
Pekoz; Berker
Kose; Selcuk
Arslan; Huseyin |
Tampa
Tampa
Tampa |
FL
FL
FL |
US
US
US |
|
|
Assignee: |
University of South Florida
Tampa
FL
|
Family ID: |
66949669 |
Appl. No.: |
16/167034 |
Filed: |
October 22, 2018 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
62623330 |
Jan 29, 2018 |
|
|
|
62609866 |
Dec 22, 2017 |
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04L 27/264 20130101;
H04B 2001/1045 20130101; H04L 27/2614 20130101; H04B 1/1036
20130101; H04B 1/1027 20130101; H04L 5/0007 20130101; H04L 25/03834
20130101; H04L 27/2607 20130101; H04L 27/2628 20130101; H04B 17/336
20150115 |
International
Class: |
H04B 1/10 20060101
H04B001/10; H04L 27/26 20060101 H04L027/26; H04B 17/336 20060101
H04B017/336 |
Goverment Interests
GOVERNMENT SUPPORT STATEMENT
[0002] This invention was made with government support 1609581
awarded by the National Science Foundation. The Government has
certain rights in the invention.
Claims
1. An adaptive windowing method for cellular communication
networks, the method comprising: determining a network normalized
received power (NNRP) for each of a plurality of links between a
transmitter and a receiver in a cellular communication; ranking the
plurality of links based upon the NNRP; maximizing transmitter
windowing for links having a higher NNRP ranking; and maximizing
receiver windowing for links having a lower NNRP ranking.
2. The method of claim 1, where the NNRP is determined using one or
more SNR values, and wherein the one or more SNR values are
selected from all users used to calculate the NNRP, only the SNR
values of a user of interest and the users utilizing adjacent bands
to the user of interest, only the NNRP of the user of interest and
the NNRP of all users of the cellular communication network.
3. The method of claim 1, further comprising aligning an alignment
signal on top of the transmitter windowing.
4. The method of claim 1, further comprising optimizing one or more
of transmitter windowing, alignment filter duration, alignment
filter coefficients, optimization weights between PAPR and OOB
emission reduction of alignment signals and receiver windowing to
maximize a capacity of the cellular communication network. The
method of claim 1, further comprising maximizing transmitter
windowing for links having a higher NNRP ranking and maximizing
receiver windowing for links having a lower NNRP ranking.
6. The method of claim 1, wherein transmitter windowing reduces
out-of-band (OOB) emissions on the cellular communication
network.
7. The method of claim 3, wherein the alignment signal reduces
out-of-band (OOB) emissions on the cellular communication
network.
8. The method of claim 3, wherein the alignment signal reduces PAPR
of the waveforms employed in the cellular network.
9. The method of claim 1, wherein receiver windowing reduces
adjacent channel interference (ACF) on the cellular communication
network.
10. The method of claim 3, wherein optimizing an alignment filter
duration of the alignment signal reduces ACI on the cellular
communication network.
11. The method of claim 3, further comprising designing the
alignment signal and an alignment filter associated with the
alignment signal, wherein designing the alignment signal and
associated alignment filter further comprises: maximizing
optimization weight of hands containing signal of links having a
higher NNRP ranking in the design of the frequency response of the
alignment filter; maximizing optimization weight of step response
linearity in the design of the alignment filter for links having a
higher NNRP ranking and maximizing optimization weight for
frequency response in the design of the alignment filter for links
having a lower NNRP ranking; and maximizing optimization weight of
OOB emission reduction in the design of the alignment signal for
links having a higher NNRP ranking and maximizing optimization
weight of PAPR reduction in the design of the alignment signal for
links having a lower NNRP ranking.
12. A system for adaptive windowing method of cellular
communication networks, the system comprising: a plurality of
transmitters and a plurality of receivers in a cellular
communication network; a base station coupled to the plurality of
transmitters and to the plurality of receivers, the base station
configured for; determining a network normalized received power
(NNRP) for each of a plurality of links between a transmitter of
the plurality of transmitters and a receiver of the plurality of
receivers in the cellular communication network; ranking the
plurality of links based upon the NNRP; maximizing transmitter
windowing for links having a higher NNRP ranking; and maximizing
receiver windowing for links having a lower NNRP ranking.
13. The system of claim 12, where the NNRP is determined using one
or more SNR values, and wherein the one or more SNR values are
selected from all users used to calculate the NNRP, only the SNR
values of a user of interest and the users utilizing adjacent bands
to the user of interest, only the NNRP of the user of interest and
the NNRP of all users of the cellular communication network.
14. The system of claim 12, wherein the base station is further
configured for aligning an alignment signal on top of the
transmitter windowing.
15. The system of claim 12, wherein the base station is further
configured for comprising optimizing one or more of transmitter
windowing, alignment filter duration, alignment filter
coefficients, optimization weights between PAPR and OOB emission
reduction of alignment signals and receiver windowing to maximize a
capacity of the cellular communication network.
16. The system of claim 12, wherein the base station is further
configured for maximizing transmitter windowing for links having a
higher NNRP ranking and maximizing receiver windowing for links
having a lower NNRP ranking.
17. The system of claim 14, wherein the base station is further
configured for designing the alignment signal and an alignment
filter associated with the alignment signal, wherein designing the
alignment signal and associated alignment filter further comprises:
maximizing optimization weight of bands containing signal of links
having a higher NNRP ranking in the design of the frequency
response of the alignment filter; maximizing optimization weight of
step response linearity in the design of the alignment filter for
links having a higher NNRP ranking and maximizing optimization
weight for frequency response in the design of the alignment filter
for links having a lower NNRP ranking; and maximizing optimization
weight of OOB emission reduction in the design of the alignment
signal for links having a higher NNRP ranking and maximizing
optimization weight of PAPR reduction in the design of the
alignment signal for links having a lower NNRP ranking.
18. One or more non-transitory computer-readable media having
computer-executable instructions for performing a method of running
a software program on a computing device, the computing device
operating under an operating system, the method including issuing
instructions from the software program comprising: determining a
network normalized received power (NNRP) for each of a plurality
between a transmitter and a receiver in a cellular communication
network; ranking the plurality of links based upon the NNRP;
maximizing transmitter windowing for links having a higher NNRP
ranking; and maximizing receiver windowing for links having a lower
NNRP ranking.
19. The media of claim 18, where the NNRP is determined using one
or more SNR values, and wherein the one or more SNR values are
selected from all users used to calculate the NNRP, only the SNR
values of a user of interest and the users utilizing adjacent bands
to the user of interest, only the NNRP of the user of interest and
the NNRP of all users of the cellular communication network.
20. The media of claim 18, further comprising computer-executable
instructions for aligning an alignment signal on top of the
transmitter windowing.
21. The media of claim 18, further comprising computer-executable
instructions for designing the alignment signal and an alignment
filter associated with the alignment signal, wherein designing the
alignment signal and associated alignment filter further comprises:
maximizing optimization weight of bands containing signal of links
having a higher NNRP ranking in the design of the frequency
response of the alignment filter; maximizing optimization weight of
step response linearity in the design of the alignment filter for
links having a higher NNRP ranking and maximizing optimization
weight for frequency response in the design of the alignment filter
for links having a lower NNRP ranking; and maximizing optimization
weight of OOB emission reduction in the design of the alignment
signal for links having a higher NNRP ranking and maximizing
optimization weight of PAPR reduction in the design of the
alignment signal for links having a lower NNRP ranking.
22. The media of claim 18, further comprising computer-executable
instructions for maximizing transmitter windowing for links having
a higher NNRP ranking and for maximizing receiver windowing for
links having a lower NNRP ranking.
Description
CROSS-REFERENCE TO RELATED APPLICATIONS
[0001] This application claims priority to U.S. Provisional Patent
Application No. 62/623,330 filed on Jan. 29, 2018, entitled
"Network-Aware Adjacent Channel Interference Rejection and Out of
Band Emission Suppression" and to U.S. Provisional Patent
Application No. 62/609,866 filed on Dec. 22, 2017, entitled
"Enhancing Performance of Beyond 5G Networks Through Adaptive
Windowing and CP Alignment, both of which are incorporated by
reference herein in their entirety.
BACKGROUND OF INVENTION
[0003] Near-far problem is one of the most critical issues in
wireless communication networks, significantly limiting network
capacity. Received power of a transmitter's signal located far from
a base station is much less than that of a nearer transmitter due
to increased propagation losses. Without precautions, the adjacent
channel interference (ACI) due to nearer transmitters can severely
diminish the signal to interference plus noise ratio (SINR) of the
far node's received signal in the uplink (UL) for frequency
division multiple accessing (TDMA) systems, resulting in the far
transmitter's signal being undetectable at the base station.
Orthogonal frequency division multiplexing (OFDM) is particularly
sensitive to this issue. Conventional OFDM transmission emits
incontrovertible energy in the out-of-band (OOB), whereas
convention reception collects energy from OOB, due to the sine
response of the rectangular pulse shape.
[0004] Many measures have been proposed to increase the far
transmitter's capacity, such as power control and strict timing and
frequency synchronization across links. However, these measures
limit network capacity and the flexibility of the system. Power
control limits the power transmitted by nearer transmitters in an
effort to reduce interference, preventing transmitters with higher
received powers from communicating at rates they would have
otherwise achieved. The strict synchronization demands require the
nodes to continuously track the synchronization signals and
precisely adjust the transmit timing and frequency, accordingly.
This continuous synchronization increases user equipment (UE) power
consumption, and the added device complexity and precision
requirements increase UE costs. Furthermore, synchronicity imposes
the same waveform with the same parameters to be used by all links
in the network, which in the case of OFDM, is the same subcarrier
spacing and cyclic prefix (CP) duration. Newer cellular
communication generations are planned to allow waveforms with
different parameters that are optimal for the link requirements,
referred to as numerologies, in adjacent bands. For example, while
low power Internet of Things (IoT) devices require smaller
subcarrier spacings to converse battery, vehicular communications
require higher subcarrier spacing and shorter symbol durations to
keep the communication running in high Doppler spreads caused by
higher speeds. Such asynchronous transmission is inherently
non-orthogonal and interference is unavoidable.
[0005] Windowing of OFDM signals is a well-studied interference
management technique in the waveform domain that has garnered
attention due to its low computational complexity. Windowing can be
performed independently at the transmitter to reduce OOB emission,
or at the receiver to reduce interference caused by communication
taking place in adjacent channels, commonly referred to as adjacent
channel interference (ACI). Techniques have been proposed utilizing
different window functions for each subcarrier at the transmitter
and receiver and derived window functions for each subcarrier that
maximizes the spectral localization and interference rejection.
[0006] Another critical problem that affects communication systems
that use the OFDM waveform is its peak-to-average power ratio
(PAPR), or the crest ratio, The PAPR is defined as the ratio of the
peak power of the analog waveform to its average power. Before the
low-power analog waveform at the output of the digital-to-analog
converter is fed to the output of the transmitter; which can be an
antenna in a wireless communication system, or a fiber-optical,
co-axial or telephone wire or another medium in a wired
communication system; it is fed to a power amplifier for
amplification of the signal. The simplified relationship between
the output voltage and the input voltage of modern power amplifiers
assumes two regions; the linear region where the gain of the
amplifier is linear if the input voltage is less than the
saturation voltage followed by the saturation region for higher
voltages. In the saturation region, the output voltage is the
maximum output voltage that is supplied by the amplifier regardless
of the input voltage, hence the one-to-one relationship between the
output and the input is no longer valid. This results in a loss of
information as the input cannot be inferred from the output
waveform. To avoid such information loss, the output of the
digital-to-analog converter is scaled with a coefficient that is
less than one prior to feeding it to the power amplifier. This
process is referred to as output back-off in the literature, and
the coefficient to preserve the one-to-one relationship decreases
as the PAPR of the waveform increases. As the waveform is scaled
with a smaller coefficient, the average power decreases, resulting
in reduced signal power at the output. Furthermore, the
relationship between the input and the output is also not linear,
even in the linear region, for practical amplifiers. The output
voltage is in fact a nonlinear function of the input voltage. This
nonlinear relationship degrades the output signal, which decreases
the signal to noise ratio (SNR) if the degradation is considered a
noise, causes subcarriers of a multicarrier signal to interfere
with one another, referred to as inter-carrier interference (WO,
and increases the OOB emission. Had the amplitude of the input
voltage been constant at all times, it would have been scaled with
the same coefficient and the output would not experience any of
these problems. Therefore, high PAPR values degrade many aspects of
the communication.
[0007] A method to reduce the PAPR and OOB emission of the OFDM
waveform involves alignment signals. Alignment signals are designed
to reduce the PAPR and OOB emission of the signal they are designed
for when added to it and are also designed to "align" with the null
space of the receiver pulse function upon convolution with the
alignment filter. Thus, they minimize the problems that are
experienced at the transmitter, and upon convolution with the
alignment filter at the receiver, they disappear and do not cause
problems to the receiver.
[0008] Current methods focus on windowing performed by extending
the symbols by an amount which is arbitrarily determined, in
addition to standard CP duration, wherein the focus is on deriving
window functions optimized according to maximizing standard
performance metrics. The currently proposed extensions reduce the
symbol rate and change the frame structure defined in the standard,
thus creating nonstandard signals that are not orthogonal to the
symbols that aims to share the same numerology. This is not
acceptable in the current cellular communication standards.
Furthermore, extending the symbol duration relentlessly causes the
symbol duration to exceed the coherence time of the channel, which
is a critical problem for high-speed vehicular communications.
Attempts have been made to improve spectral efficiency of windowed
OFDM systems by utilizing less extension for inner subcarriers and
assigning edge subcarriers to users with lower delay spread to use
more windowing in edge subcarriers. However, this scheme still dos
not comply with the standard frame structure. Additionally,
standard compliant schemes have been derived for receiver windowing
durations that optimize reception of each subcarrier in the case in
which intersymbol interference (ISI) and ACI occur simultaneously
and pulse shapes of transmitters operating in adjacent bands cannot
be controlled, in the absence of any extension designated for
windowing. However, determining whether it is more beneficial to
window a duration at the transmitter or receives has not been
previously addressed. Current methods focusing on alignment signals
use part of the standard CP extension assuming it is not disrupted
by anything else, yet again there's no study on the amount of CP to
be assigned for alignment signals, especially in combination with
transmit and receiver windowing, to optimize network conditions.
Furthermore, PAPR and OOB emission are competing goals in the
calculation of the alignment signals, and the earlier studies
assign weights to each goal randomly. No study has been made to
personalize the weights depending on the user's and the network's
condition. The design of the filters themselves was not studied
either.
[0009] Accordingly, what is needed in the art is an improved system
and method that optimizes the combination of transmitter windowing,
receiver windowing and alignment signals to maximize the overall
capacity of the communication network.
SUMMARY OF THE INVENTION
[0010] In various embodiments, the present invention provides
capacity gains if transmitter and receiver windowing are utilized
jointly and adaptively to minimize network interference and
maximize network capacity. The invention additionally proposes
metrics that can be used to guess the optimum window durations
based on the received power of links in adjacent bands, thereby
eliminating the need for network-wide optimization.
[0011] In one embodiment, the present invention provides, an
adaptive windowing method for cellular communication networks which
includes, determining a network normalized received power (NNRP)
for each of a plurality of links between a transmitter and a
receiver in a cellular communication network, ranking the plurality
of links based upon the NNRP, maximizing transmitter windowing for
links having a higher NNRP ranking and maximizing receiver
windowing for links having a lower NNRP ranking.
[0012] In the present invention, the network normalized received
power (NNRP) for each of the plurality of links is based at least
upon a ratio of received power of a link to that of adjacent links
and the method further includes optimizing transmitter windowing
and receiver windowing to maximize a capacity of the cellular
communication network.
[0013] In an additional embodiment, the present invention provides
a system for adaptive windowing method of cellular communication
networks comprising a plurality of transmitters and a plurality of
receivers. The system further includes a base station coupled to
the plurality of transmitters and to the plurality of receivers,
the base station configured for determining a network normalized
received power (NNRP) for each of a plurality of links between a
transmitter of the plurality of transmitters and a receiver of the
plurality of receivers in the cellular communication network,
ranking the plurality of links based upon the NNRP, maximizing
transmitter windowing for links having a higher NNRP ranking and
maximizing receiver windowing for links having a lower NNRP
ranking.
[0014] in another embodiment, the present invention provides, one
or more non-transitory computer-readable media having
computer-executable instructions for performing a method of running
a software program on a computing device, the computing device
operating under an operating system. The method includes issuing
instructions from the software program comprising, determining a
network normalized received power (NNRP) for each of a plurality of
links between a transmitter and a receiver in a cellular
communication network, ranking the plurality of links based upon
the NNRP, maximizing transmitter windowing for links having a
higher NNRP ranking; and maximizing receiver windowing for links
having a lower NNRP ranking.
[0015] Accordingly, the present invention provides an improved
system and method which combines adaptive transmitter windowing,
receiver windowing and alignment signals based upon network
normalized received power (NNRP) for each of a plurality of links
between a transmitter and a receiver in a cellular communication
network.
BRIEF DESCRIPTION OF FIGURES
[0016] FIG. 1A is a visual demonstration of the adaptive cyclic
prefix (CP) methodology, in accordance with an embodiment of the
present invention.
[0017] FIG. 1B is a visual demonstration of the standard symbol
structure and the symbol structure proposed by previous windowing
schemes in addition to the simplified and optimized windowing
scheme of adaptively using the readily available CP for windowing
in either transmitter or receiver side depending upon the power
offset, without utilizing alignment signals, as proposed by the
present invention.
[0018] FIG. 2 is an illustration of a conventional OFDM-based
system on which various methods of the present invention are built
upon.
[0019] FIG. 3A is an illustration of the transmitter of the
advanced transmitter and receiver windowing OFDM coupled with
alignment signals. The IFFT operation of the conventional
transmitter is performed within transmit pulse shaping in the
transmitter windowing implementation.
[0020] FIG. 3B is an illustration of the receiver of the advanced
transmitter and receiver windowing OFDM coupled with alignment
signals. The ITT operation of the conventional receiver is
performed within receive pulse shaping in the receiver windowing
implementation.
[0021] FIG. 4 is a graphical illustration of the percent capacity
gain compared to no windowing if whole undisturbed CP duration is
used either for receiver windowing, transmitter windowing, or
adaptively as the proposed initial guess of the optimization for
the links with lowest and highest relative received power and the
mean of all links. Alignment signals are not utilized in obtaining
the results presented in this figure, and the initial guess refers
to the case where transmitter windowing durations are equal to
"guess" values presented in FIG. 6, with complementary receiver
windowing durations.
[0022] FIG. 5 is a graphical illustration of a percent capacity
gain of the method of the present invention over a no-windowing
condition if the entire clean CP duration is used either for
receiver or transmitter windowing, solely, adaptively as guessed
and optimized, for LTE normal and extended CP, of the links with
the lowest and highest NNRP and the network average. Alignment
signals are not utilized in obtaining the results presented in this
figure, and the guess refers to the case where transmitter
windowing durations are equal to "guess" values presented in FIG.
6, with complementary receiver windowing durations.
[0023] FIG. 6 is a graphical illustration of initial guesses,
optimum values and best-fit of the optimum values of the ratio of
clean CP utilized for transmitter windowing to total clean CP
duration, as a function of NNRP for extended CP duration. Alignment
signals are not utilized in obtaining the results presented in this
figure.
[0024] FIG. 7 is an illustration demonstrating how the transmitter
windowed samples are generated by overlapping scaled CPs and CSs of
consecutive OFDM symbols of which indices are given in the
subscripts, in accordance with an embodiment of the present
invention
[0025] FIG. 8 is an illustration demonstrating how indexing and
identification of CP and symbol parts are used in a receiver
windowing operation, in accordance with an embodiment of the
present invention.
[0026] FIG. 9 is a graphical illustration of the optimum transmit
(T) and receive (R) window durations of a two user network as a
function of the power offset between them. Alignment signals are
not utilized in obtaining the results presented in this figure.
[0027] FIG. 10 is a graphical illustration of the BER of the
outpowering user as a function of the power offset for various
windowing algorithms. The rectangular algorithm refers to the
transceiver described in FIG. 2, whereas proposed (simplified) and
proposed (optimum) algorithms refer to the algorithms presented in
FIG. 1B, Alignment signals are not utilized in obtaining the
results presented in this figure.
[0028] FIG. 11 is a graphical illustration of the BER of the
outpowered user as a function of their SNR for various windowing
algorithms. The rectangular algorithm refers to the transceiver
described in FIG. 2, whereas proposed (simplified) and proposed
(optimum) algorithms refer to the algorithms presented in FIG. 1B.
Alignment signals are not utilized in Obtaining the results
presented in this figure.
[0029] FIG. 12 is a graphical illustration of the proportional fair
network spectral efficiency for various windowing algorithms as a
function of the outpowered user's SNR. The rectangular algorithm
refers to the transceiver described in FIG. 2, whereas proposed
(simplified) and proposed (optimum) algorithms refer to the
algorithms presented in
DETAILED DESCRIPTION OF THE INVENTION
[0030] The present invention addresses how network capacity can be
further improved if the pulse shapes and alignment signals of the
transmitters can be coordinated while conserving the standard frame
structure, that is, not adding any additional extensions other than
CP and using only the present CP for windowing and alignment
signals. It is proposed that transmitter and receiver windowing and
alignment signals be implemented simultaneously to maximize fair
proportional network spectral efficiency with the goal of
determining the amount of windowing that should be applied at
either side, for each particular use. In various embodiments, the
present invention illustrates that the transmitter and receiver
windowing duration of each user maximizing the network spectral
efficiency can be effectively guessed using the power offset
between the user of interest and the users employing adjacent
bands. It is also shown that the user with highest received power
is the prominent source of ACI impacting other users, whereas the
ACI caused by lower powered users utilizing adjacent band has
little effect on its spectral efficiency. Therefore, the user with
highest receive power must window most of the CP duration at the
transmitter for the sake of maximizing the fair proportional
network spectral efficiency. Furthermore, while transmitter
windowing does not seem to have any obvious benefit to the user
applying it, even without alignment signals, it has been shown that
even for slower vehicular speeds, such as 30 km/h, the reduction in
intercarrier interference (ICI) improves the capacity of the user
applying transmitter windowing. On the other hand, reducing ACI
caused by the user with the lowest received power was the veriest
effect on other users' spectral efficiencies, while the capacity of
the user with the lowest received power is limited heavily by the
ACI cause by users with higher received power utilizing adjacent
bands. Thus, the user with the relatively lowest received power
must window most of the clean CP at the receiver and focus on
improving their reception, rejecting as much ACI as possible. For
nonextreme power offset values, the portion that is adopted for
transmitter and receiver windowing can be determined as a function
of the power offset.
[0031] In various embodiments, the present invention provides a
system and method for utilizing windowing and CP alignment based on
relative received powers of links in a way that it is beneficial
for all nodes in the cell. The link with highest relative received
power is the prominent source of OOB leakage, and the OOB leakage
of lower powered links transmitting in adjacent bands has little
effect on its capacity, which is bounded by PAPR-related
distortions. Therefore, the link with highest relative received
power windows most of the clean CP at the transmitter, benefiting
from longer alignment signal (AS) duration resulting in maximum
PAPR reduction, maximizing its capacity. In this context, dirty and
clean CP refers to the CP portions that are and are not disturbed
by multipath interference, respectively.
[0032] Furthermore, the OOB reduction provided by the long
transmitter window duration as well as the alignment signal
improves the capacity of the rest of the network. The weight for
the OOB emission reduction in AS design is higher for the
transmitter with the highest network normalized received power to
help aid this purpose. Reducing OOB leakage of the link with lowest
relative received power, on the other hand, has the veriest effect
on other link's capacities, while the capacity of the link with
lowest relative received power is limited by the OOB leakage of
higher relative received power links operating in adjacent bands.
Thus, the link with lowest relative received power uses most of the
clean CP duration for receiver windowing, rejecting as much OOB
energy as possible, maximizing its capacity. The weight for the
PAPR reduction in AS design is increased for the links with lower
network normalized received power so that these links further
improve their own capacity, which in turn increases the fair
proportional network capacity. The links with intermediate relative
received powers adopt a portion of clean CP for transmitter
windowing, and the remainder for receiver windowing, depending on
the ratio of their received power to adjacent links received power.
The optimization weights of AS design shift as well from PAPR
reduction to OOB emission reduction as relative received power
increases from the network minimum to network maximum.
[0033] The general scheme 100 of the present invention is visually
demonstrated in FIG. 1. As shown in FIG. 1A, the present invention
proposes utilizing transmitter windowing and receiver windowing,
simultaneously, to maximize network capacity. Generally, the
relative received power of a link 105 increases as the distance to
the base station (BS) decreases. In the present invention, as the
relative received power of a link 105 increases, transmitter
windowing 120 increases, hence allowing increased filtered
alignment signal duration 115, and receiver windowing 125
decreases. As such, the alignment filter 135 of the link is
adjusted, based upon received powers of links and distances to the
base station in a way that is beneficial to all links in the
cellular communication network. The order of the alignment filter
135 may also depend on the relative received power of a link 105.
The CP 130 includes both CP portions that are, and are not,
disturbed by multipath interference 140. The relative received
power 105 may be calculated by a base station using the received
powers of all links in the network or may be calculated for each
link depending solely on the received power of the links employing
adjacent bands only, either by the base station or by the
transmitting links themselves. The CP is used to implement
transmitter windowing, receiver windowing and alignment signal
operations as well as its initial design goal to accommodate
multipath propagation. The duration allocated to implement the
transmitter windowing operation increases, whereas the duration
allocated to implement receiver windowing decreases as the user's
power compared to the rest of the network increases. Another key
difference, previously unknown in the art, is that the alignment
signal is aligned on top of the transmitter window. Transmitter
windowing operation does not benefit the user doing the operation
and is performed solely to benefit the network. By allowing the
alignment signal to align on top of the transmitter window, the
duration allotted to transmitter windowing enables more effective
alignment signals by augmenting the null space. This converts
transmitter windowing to a technique that also benefits the user
applying it.
[0034] The standard symbol structure and the symbol structure
proposed by previous windowing work is shown in FIG. 1B, along with
the simplified and optimized idea of adaptively using the readily
available CP for windowing in either side, depending upon the power
offset, as proposed by the present invention. This figure
represents a partial algorithm that only utilizes joint transmitter
and receiver windowing and does not include the alignment signals.
If no alignment signals are utilized, the transmitter and receiver
windowing operations are not exclusive, and the tails of the
windows can overlap, which was also not possible with previous art.
As shown in FIG. 1B, the width of the rectangles represent allowed
times for the actual OFDM symbol, CP, and further cyclic extensions
for "T" transmitter and "R" receiver windowing, while dashed
overlays and the round dot overlays demonstrate transmitter and
receiver windowing of the underlying area, respectively. Previous
windowing techniques known in the art affix additional extensions,
thus breaking the standard symbol structure and they do not focus
on the amount of these extensions and when they must be used. In
contrast, the implementation of the windowing scheme of the present
invention does not change the standard structure and focuses on how
the readily available extension should be used as a function of the
power offset between user communicating in adjacent bands.
Numerical results confirm that fair proportional network spectral
efficiency can be increased greatly without disrupting the standard
frame structure by utilizing CP adaptively, and that power offset
between user utilizing adjacent bands is a clear metric in
determining the side to apply windowing.
[0035] As shown with reference to FIG. 2, part of the method of the
present invention may be employed in system 200 comprising an OFDM
transmitter 205 and/or an OFDM receiver 225. As shown with
reference to FIG. 2, the OFDM transmitter 205 includes a modulation
module 210 configured to receive incoming data bits and to generate
digital modulated symbols comprising a plurality of symbols. The
OFDM transmitter 205 additionally includes an Inverse Fast Fourier
Transform (IFFT) module 215 to receive the plurality of modulated
symbols from the modulation module 210. The IFFT module 215
receives incoming symbols to generate an OFDM-based signal
comprising a plurality of subcarriers. The IFFT module 215 operates
as a transmitter filter to filter the subcarriers of the OFDM-based
signal using the proposed subcarrier specific based windowing
scheme to generate a filtered OFDM-based signal. The filtered
OFDM-based signal is then provided to a digital-to-analog module
220 of the transmitter prior to transmission of the filtered
OFDM-based signal over the channel to a base station (not shown),
In addition, the OFDM receiver 225 includes an analog-to-digital
module 230 configured to receive incoming OFDM-based signals
comprising a plurality of subcarriers that has been transmitted
over the channel from the base station. The analog-to-digital
module 230 provides the digital representation of the OFDM-based
signals to a Fast Fourier Transform (FFT) module, operating as a
receiver filter 235 to filter the subcarriers of the OFDM-based
signal using the proposed subcarrier specific windowing scheme of
the present invention to generate a filtered OFDM-based signal, The
filtered OFDM-based signal is then provided to a demodulation
module 240 of the receiver.
[0036] The whole method of the present invention is employed in a
system comprising a transmitter and receiver windowed OFDM with
alignment signal transmitter 300 and a transmitter and receiver
windowed OFDM with alignment signal receiver 400. The transmitter
and receiver windowed OFDM with alignment signal transmitter 300 is
illustrated in FIG. 3A and the transmitter and receiver windowed
OFDM with alignment signal receiver 400 is illustrated in FIG.
3B.
[0037] With reference to FIG. 3A, in one embodiment, the SNR of a
desired user of interest 305, the SNR of a first adjacent user 310
and the SNR of a second adjacent user 315 may be provided to the
window and alignment filter duration calculation module 320 to
obtain the optimum transmitter window duration, the optimum
receiver window durations 330 and the optimum alignment filter
duration 335 of the user of interest. In this embodiment, the SNR
of the user of interest 305 and the SNR of two adjacent users 310,
315 are used in the calculation of the window and alignment filter
duration, however in alternate embodiments, the NNRP of the user of
interest may be used for the calculation or the SNRs of all the
users that are used to calculate the network NNRPs may be used to
perform the calculation.
[0038] In this embodiment, the resulting alignment filter duration
335, the tone allocation 340 of the user of interest, the SNRs of
the user in interest and the users utilizing adjacent bands are
provided to the alignment filter design module 345 for the user of
interest. This module 345 may be designing an optimum finite
impulse response (FIR) filter from the values given; or a book of
FIR filter coefficients may be designed before the device was
manufactured and be recorded on the device on an electronic medium,
and the module 345 may be choosing the filter that best fits the
values provided to it. When calculating the optimum FIR filter
coefficients, the optimization used in module 345 may give more
importance to suppressing the interference coming from the adjacent
user with higher SNR. In various embodiments, when calculating the
optimum FIR filter coefficients, the optimization used in module
345 may also take into account the NNRP of the user in interest, or
the SNRs of all users that are used to calculate the network NNRPs,
or the SNRs of only the user in interest and the users utilizing
adjacent bands to calculate coefficients with a more uniform
powered impulse response profile, as filters with such traits
provide better PAPR suppression, if the desired user is outpowered.
In the case of optimum FIR filter design, the filter coefficients
350 must be submitted to the transmitter and receiver windowed OFDM
with alignment signal receiver 400 through a separate communication
channel, or in the case of using a predefined filter, the index of
the filter from the filterbook must be submitted to the transmitter
and receiver windowed OFDM with alignment signal receiver 400 of
FIG. 3B. At the same time, the bits of the desired user 360 are
modulated at a modulation module 365 using any desired digital
modulation technique, which may be PSK, QAM or APSK. The resulting
symbols are then modulated to tones allocated to the user of
interest by filtering them with the transmit pulse shaping filters
of each tone by the transmit pulse shaping module 370. The transmit
pulse shaping module 370 calculates the transmit pulse shape of
each tone by convolving the Fourier transform matrix column
associated with that tone, extended periodically by the GP duration
minus the transmitter window duration, with the discrete prolate
spheroidal sequence associated with that tone for the provided
transmitter windowing duration. The samples obtained at the output
of the transmit pulse shaping module 370 and the alignment filter
design module 345 and the SNRs of desired and users adjacent to
desired 305, 310, 315 are provided to the alignment signal design
module 355 along with the calculated receiver window duration 330
so that the alignment signal is designed. The alignment signal
design module 355 takes into account the SNRs of desired and users
adjacent to desired 305, 310, 315 to determine the weight between
PAPR suppression or OOB emission for the user in interest. The
receiver window duration 330 reveals the null space of the receiver
pulse shape. The designed alignment signal 375 is then added to the
transmit samples 380 and the resulting samples 385 are fed to the
digital-to-analog converter module 390 for transmission over the
transmission medium 395.
[0039] With reference to FIG. 3B, after propagating through the
transmission medium, combined with the signals transmitted for
other links, which are obtained using a similar procedure, the
waveform 395 arrives at the transmitter and receiver windowed OFDM
with alignment signal receiver 400. The received waveform is
quantized and sampled by the analog-to-digital conversion module
405. The digital signal is then filtered 410 with the alignment
filter coefficients 350 at a convolve with alignment filter module
410, wherein the alignment filter coefficients 350 are the same
coefficients used by the transmitter 300. Time and frequency
synchronization is performed to the filtered samples using the
synchronization module 415 using known symbols implemented in the
transmitted waveform, the tone allocation 340 and the receiver
window duration 330 from the transmitter 300. The alignment filter
coefficients 350 and the receiver window duration 330 may be
communicated to the receiver from another communication channel, or
the receiver may also have a duplicate window and alignment filter
duration calculation module 320 and alignment filter design module
345 and calculate these values using its knowledge of 305, 310,
315. Receiver pulse shaping 420 is applied to synchronized samples,
where the receiver pulse shape of each tone allocation 340 is
obtained by convolving the discrete prolate spheroidal sequence
corresponding to that tone with the Fourier matrix row
corresponding to that tone. Channel is estimated by the channel
estimation block 425 using the received symbols and known symbols
located within data symbols, similar to synchronization 415, and
unknown channel coefficients between estimated channel coefficients
are estimated from the estimated channel coefficients. The received
symbols 435 and channel coefficients 440 are fed to the
equalization block 430 to remove the effect of the channel from the
received symbols 435. The equalized symbols are demodulated at a
demodulation module 445 to obtain the received bits 450 at the
receiver 400.
[0040] In accordance with one exemplary embodiment of the system
and method of the present invention, links are given indices
u.di-elect cons.{1, 2, . . . , U} in the order they are utilizing
the spectrum. Pilot tones are used for synchronization and channel
estimation to better demonstrate the gains of reduced interference
on capacity. In the following discussion, (.cndot.).sup.T and
(.cndot.).sup.H denote the transpose and Hermitian operations,
0.sub.axb and 1.sub.axb denotes matrices or zeros and ones with a
rows and b columns, diag (v) returns a square diagonal matrix with
the elements of vector v on the main diagonal, (.mu.,
.sigma..sup.2) represents complex Gaussian random vectors with mean
.mu. and variance .sigma..sup.2, and toep(c, r) corresponds to the
Toeplitz matrix of which first column is c and first row is r.
Wherein the uth link's base time domain samples .XI..sub.u.di-elect
cons.
K u + N u + K T u .times. I u ##EQU00001##
are obtained using:
.XI..sub.u=.SIGMA..sub.m=1.sup.M.sup.uA.sub.u,m(Q.sub.u,m+D.sub.u,m)
(1)
where Q.sub.u,m.di-elect cons..sup.1.times.I.sup.u and
D.sub.u,m.di-elect cons..sup.1.times.l.sup.u are row vectors
consisting of pilot and data symbols to be transmitted by uth
link's mth subcarrier, respectively, M.sub.u is the number of
subcarriers employed by uth link in each OFDM symbol, I.sub.u is
the number of OFDM symbols the uth link transmits in a packet,
N.sub.u is the fast Fourier transformation (FFT) size utilized by
uth links transmitter, F.sub.u,m.di-elect cons..sup.1.times.N.sup.u
is a row of the normalized N.sub.u-point FFT matrix corresponding
to the mth subcarrier of uth user, and the CP (cyclic prefix)
addition and transmitter windowing matrix A.sub.u,m.di-elect
cons.
K u + N u + K T u .times. N u ##EQU00002##
shown in (2) uses uth link's mth subcarrier's transmitter window's
ramp-up tail coefficients w.sub.T.sub.u,m.di-elect cons.
K T u , ##EQU00003##
where K.sub.u is the total number of uth link's CP samples and
K.sub.T.sub.u is the number of ramp-up tail coefficients of the
transmitter window function.
A u , m = [ O K T u .times. K T u O K T u .times. N u - K u - K T u
diag ( w T u , m ) O K T u .times. K u - K T u O K u - K T u
.times. K T u O K u - K T u .times. N u - K u - K T u O K u - K T u
.times. K T u I K u - K T u I K T u O K T u .times. N u - K u - K T
u O K T u .times. K T u O K T u .times. K u - K T u O N u - K u - K
T u .times. K T u I N u - K u - K T u O N u - K u - K T u .times. K
T u O N u - K u - K T u .times. K u - K T u O K T u .times. K T u O
K T u .times. N u - K u - K T u I K T u O K T u .times. K u - K T u
O K u - K T u .times. K T u O K u - K T u .times. N u - K u - K T u
O K u - K T u .times. K T u I K u - K T u diag ( 1 K T u .times. 1
- w T u , m ) O K T u .times. N u - K u - K T u O K T u .times. K T
u O K T u .times. K u - K T u ] ( 2 ) ##EQU00004##
[0041] Ramp-up and ramp-down transmitter window tails of
consecutive symbols are overlapped to form the transmit windowed
samples matrix X.sub.u.di-elect
cons..sup.N.sup.u.sup.+K.sup.u.sup..times.I.sup.u.sup.+1:
X u = [ I K u + N u O K u + N u .times. K T u ] .times. [ .XI. u O
K u + N u .times. K T u .times. 1 ] + [ O K T u .times. K u + N u I
K T u O K u + N u - K T u .times. K u + N u O K u + N u - K T u
.times. K T u ] .times. [ O K u + N u .times. K T u .times. 1 .XI.
u ] ( 3 ) ##EQU00005##
[0042] The alignment filter coefficients of uth link,
g.sub.u.di-elect cons.
K A u .times. 1 ##EQU00006##
are used to construct the alignment convolution matrix
G.sub.u.di-elect
cons..sup.(K.sup.u.sup.+N.sup.u.sup.).times.(K.sup.u.sup.+N.sup.u.sup.),
where
G u = toep ( [ g u ; 0 ( K u + N u - K A u ) .times. 1 ] , [ g u (
1 ) , 0 1 .times. ( K u + N u - 1 ) ] ) . ##EQU00007##
If B.sub.u,m.di-elect
cons..sup.N.sup.u.sup..times.N.sup.u.sup.+K.sup.u is the nth link's
CP removal and receiver windowing matrix used in the reception of
the mth subcarrier shown in (6).
B u , m = [ O N u - K R u .times. K u - K R u O N u - K R u .times.
K R u I N u - K R u O N u - K R u .times. K R u O K R u .times. K u
- K R u diag ( w R u , m ) O K R u .times. N u - K R u diag ( 1 K R
u .times. 1 - w R u , m ) ] ( 6 ) ##EQU00008##
[0043] Where K.sub.R.sub.u is the duration of the receiver window
ramp and
w R u , m .di-elect cons. K R u ##EQU00009##
is the uth link's mth subcarrier's receiver window's ramp-up tail
coefficients. Then the alignment signal precoder
P u .di-elect cons. ( N u + K u ) .times. ( K u - K R u )
##EQU00010##
is obtained as P.sub.u=ker(B.sub.m,uG.sub.u) for any m. Then uth
link's alignment signal matrix
S u * .di-elect cons. ( K u - K R u ) .times. I u ##EQU00011##
is calculated accordingly as
S u * = arg min S u ( 1 - .lamda. u ) F O ( X u + P u S u ) 2 +
.lamda. u X u + P u S u .infin. ##EQU00012##
subject to B.sub.m,uG.sub.uP.sub.uS.sub.u for any m and
.parallel.P.sub.uS.sub.u.parallel..sub.2.ltoreq. {square root over
(.alpha.)}.parallel.X.sub.u.parallel..sub.2, where F.sub.O consists
of the rows of the N.sub.u-point Fourier transform matrix for which
frequencies contain signals of other users and .alpha. is a
parameter that limits the power of the alignment signal, and
.lamda..sub.u.di-elect cons.[0,1] is a weighting factor for the
joint optimization of OOB emission and PAPR.
[0044] The transmit sample sequence t.sub.u is obtained by
converting T.sub.u=X.sub.u+P.sub.uS.sub.u.sup.* from parallel to
series. In practical systems, the sample sequence is then converted
to an analog waveform, which is amplified using a power amplifier
(PA) experiencing amplitude-amplitude and amplitude-phase
distortions as is known in the art.
[0045] The received samples are given by:
[0046] r=n+.SIGMA..sub.u=1.sup.U[0.sub.d.sub.u
.gamma..sub.u(h.sub.u*t.sub.u)] where
n.about.(0,.sigma..sub.n.sup.2) is background additive white
Gaussian Noise (AWGN) of which .sigma..sub.n is defined for various
types of user equipment (LTE) and base stations (BW),
h.sub.u.di-elect cons..sup.L.sup.u is the Rayleigh fading channel
coefficients of uth user. To align the alignment signal to the
kernel of the receive pulse shape of uth user, the received samples
are convolved with the alignment filter of uth user:
r.sub.u=g.sub.u*r. As the propagation delay is not known at the
receiver, packet timing offset estimation is performed using the
cross-correlation between r and transmitted pilot samples .sub.u,
where .sub.u is obtained similar to (3) where Q.sub.u is used
instead of (Q.sub.u+D.sub.u) in (1) and converted from parallel to
series. R.sub.u .di-elect
cons..sup.N.sup.u.sup.+K.sup.u.sup..times.I.sup.u is obtained by
converting the series synchronized samples to I.sub.u parallel
streams of N.sub.U+K.sub.u samples each. Received symbols of the
mth subcarrier are obtained using:
Y.sub.u,m=F.sub.u,mB.sub.u,mR.sub.u (5)
[0047] Channel frequency response (CFR) coefficients at pilot tones
are estimated by H.sub.u=Y.sub.uOQ.sub.u for nonzero elements of
Q.sub.u. Remaining CFR coefficients, in-between known coefficients,
are interpolated using natural neighbors and the remainder are
linearly extrapolated based upon boundary gradients. Finally, data
symbols are equalized with zero forcing equalization: {circumflex
over (D)}=Y.sub.uOH.sub.u.
[0048] The present invention proposes the estimation and
utilization of K.sub.T=[K.sub.T.sub.1 K.sub.T.sub.2 . . .
K.sub.T.sub.U], K.sub.A=[K.sub.A.sub.1 . . . K.sub.A.sub.U],
K.sub.R=[K.sub.R.sub.1 K.sub.R.sub.2 . . . K.sub.R.sub.U], g.sub.u
and .lamda..sub.u .A-inverted.u values that maximize the fair
proportional network capacity. Mean channel capacity can be
calculated in idealized scenarios using background noise level,
received power estimates and expected interference, if hardware
effects are not considered. In the various exemplary embodiments,
optimum values using realistic simulations that include hardware
impairments through numerical methods and present novel metrics
that yield near-optimum window lengths and capacity gains, thereby
eliminating the need for any optimization process.
[0049] Therefore, a routine is proposed that consists of making a
heuristic initial guess using the power offset across users, then
attempting to converge to the optimal solution, iteratively. The
heuristic initial guess is performed in a centralized way by a
control unit that monitors the whole coordinated spectrum, such as
the base station (BS). In practical implementations, the iterations
can also be performed in a centralized way by the same control unit
using evolutionary algorithms, or alternatively, the iterations can
be performed independently by all nodes in a decentralized manner
using a game theoretical approach. Assuming received powers vary
slowly over time, evolving optimal lengths also improves robustness
against channel variations and evolutionary tracking reduces
computational complexity compared to solving nondeterministic
polynomial (NP)-complete optimization problems for each resource
block.
[0050] The heuristic initial guess of the transmitter window length
assumes that K.sub.u is fixed. The ratio of received power of the
desired link to that of adjacent links is obtained as
.rho..sub.u=log(.gamma..sub.u/ {square root over
(.gamma..sub.u-1.gamma..sub.u+1)}) at the intended receive of the
uth link and shared with the base station (BS) if necessary, i.e.,
the algorithm can be used for networks in which adjacent uplink
(UL), downlink (DL) and sidelink (SL) communication takes place.
The BS calculates the uth user's network normalized received power
(NNRP) by g.sub.u=.rho..sub.u-min(.rho.)/max(.rho.)-min(.rho.),
where .rho.=[.rho..sub.1 .rho..sub.2 . . . .rho..sub.U]. The
transmitter window length guesses are obtained as {circumflex over
(K)}.sub.T.sub.u=round(.beta..sub.K.sub.u) so that the user with a
higher relative received power uses more CP duration for
transmitter windowing, where .beta..sub.K.sub.u is a coefficient
depending on K.sub.u. The alignment filter duration and alignment
filter coefficients, as well as .lamda. values are also determined
based on .rho. or P values. The remaining clean CP duration is
utilized for receiver windowing, {circumflex over
(K)}.sub.R.sub.u=K.sub.u-K.sub.A.sub.u-{circumflex over
(K)}.sub.T.sub.u. Coefficients .beta..sub.K, for standard K values,
are estimated by fitting optimum window lengths obtained from
training simulations, and test simulations yield that the values
are robust. Optimum window lengths K.sub.T.sup.* and K.sub.R.sup.*
are obtained by solving the following combinatorial optimization
problem:
K T * , K R * = arg K T , K R where C u = M u 2 .DELTA. f u N u + K
u log ( 1 + 1 { Q u .times. ( R u ( Q u .noteq. 0 ) ) H } ) ( 7 )
##EQU00013##
[0051] is the uth link's capacity, where .DELTA.f.sub.u is the uth
link's subcarrier spacing.
[0052] The proposed guesses and optimized values for transmitter
and receiver windowing is implemented and compared with systems
utilizing no windowing, and with systems utilizing whole clean CP
duration for transmitter and receiver windowing for Long Term
Evolution (LTE) normal (K.sub.u=N.sub.u.sub.9/128) and extended
(K.sub.u=N.sub.u.sub.32/128) CP duration.
[0053] In an exemplary embodiment, the system parameters are as
follows: links 1 and 3, out of a total of U=4, are simulated to be
Internet of Things (IoT) UL links having .DELTA.f=15 kHz and each
utilize M=48 subcarriers during I=14 OFDM symbols generated by
performing N=256-point FFT, whereas links 2 and 4 are simulated to
be enhanced-mobile broadband (eMBB) UL links having =30 kHz and
each utilize M=24 subcarriers during I=28 OFDM symbols generated by
performing N=128-point FFT. IoT and eMBB devices employ 8 and 14
bit digital-to-analog converters (DACs), respectively, and the BS
employs a 16-bit analog-to-digital converter (ADC). LTE cell
specific reference signals are used for synchronization and channel
estimation. There is 120 kHz fixed guard band between each link.
All transmitter and receiver window coefficients are calculated
using the per-subcarrier approach. R=2 km and maximum output power
of the nodes are determined based upon their device type. All nodes
have a mobility of 60 km/h. The optimum values are obtained using
an integer genetic algorithm.
[0054] The capacity gains of links over no windowing is given in
FIG. 4. It can be seen that the link with lowest relative received
power benefits more from receiver windowing compared to transmitter
windowing, and the opposite applies to the link with the highest
relative received power, proving the reasoning of the present
invention. However, it can be seen that overall network capacity,
along with the capacity of the link with the lowest relative
received power is maximized even when window lengths are per the
calculated initial guess.
[0055] The capacity gains over no windowing are illustrated in FIG.
5. It can be seen that adaptive windowing guesses outperform
windowing whole clean CP exclusively at the transmitters or the
receiver in terms of overall network capacity and the capacity of
the link with the lowest relative received power. Mean network
capacity gained by using guesses and optimum window durations are
22% and 28%, respectively, for normal CP, and 15% and 16%,
respectively, for extended CP. One interesting observation is that
although the highest network capacity is achieved using normal CP,
the lowest NNRP links further benefit from extended CP, due to
improved interference management. This effect is not preserved in
the network level as the capacity loss due to extended effective
symbol duration outweighs the benefits of reduced ACI for higher
power users.
[0056] FIG. 6 is a scatter-plot illustrating the ratio of guesses
and optimum transmitter windowed CP to the total extended clean CP
against NNRP, as well as the linear polynomial fit of the optimum
points. The fits yield adjusted R.sup.2 values of 0.9938 and 0.9996
for normal and extended CP, respectively
.beta. 9 N 128 and .beta. 32 N 128 ##EQU00014##
were estimated as 0.88 and 0.674 from the results of the training
period of 64 random networks (with 4 users in each, a total of 256
values were obtained) and used in the guessing of initial values
for the test set, while estimates obtained from the optimal values
of the independent test set of 64 random networks are 0.8571 and
0.6731, respectively.
[0057] It has been shown that the capacity of lower NNRP links is
limited by interference of higher power links. Links with lower
power thus benefit more from receiver windowing. Links with higher
NNRP experience relatively less interference, thus should focus
more on reducing their emission by performing transmitter windowing
with most of their clean CP for the sake of the network. However,
as clean CP becomes abundant, the network capacity gain from
reduced OOB emissions, resulting from further increasing
transmitter window durations, diminishes and the network benefits
more from increased capacity of the high powered users, yielding an
inversely proportional relationship between .beta. and K. NNRP
proves to be an effective metric in guessing window lengths
maximizing network capacity and the links with the lowest NNRP.
Links with the lowest relative received powers benefit from better
interference management if extended CP lengths are utilized,
however the network average reduces due to increased effective
symbol duration.
[0058] Hardware distortions are the factor limiting the capacity of
high relative received power links. Utilizing transmitter windowing
becomes more beneficial than receiver windowing for links with
higher relative received powers. Capacity of lower relative receive
power links is limited by interference, so links with lower
relative received power benefit more from receiver windowing. Ratio
of the received power of the desired signal to that of interfering
signals is an effective metric in guessing the initial window
lengths. Applying transmitter and receiver windowing by considering
these ratios increases the average capacity and energy efficiency
of the network, as well as that of the links with the lowest
relative receive powers.
[0059] In an additional exemplary embodiment of the system and
method of the present invention, it is assumed that there are U
transmitters sharing a bandwidth B using a transmitter and receiver
windowed OFDM symbol system. In this discussion of this exemplary
embodiment, (.cndot.).sup.T, (.cndot.).sup.* and (.cndot.).sup.H
denote the transpose, conjugate and Hermitian operations, A [a, b]
is the element in the ath row and bth column of matrix A,
A.circle-w/dot.B and AOB correspond to Hadamard multiplication and
division of matrices A and B and A by B, A.sup..circle-w/dot.2
refers to A.circle-w/dot.A.sup.*. 0.sub.a.times.b and
1.sub.a.times.b denotes matrices or zeros and ones with A rows and
B columns, (.mu., .sigma..sup.2) represents complex Gaussian random
vectors with mean .mu. and variance .sigma..sup.2.
[0060] In this embodiment, it is assumed that one node, referred to
as the next generation Node B (gNB), aims to receive the
information transmitted by all transmitters correctly, and all
transmitters intend to convey information to this node; a situation
that commonly arises in UL reception. Each transmitter u samples
this band using an N.sub.u- point Fast Fourier Transformation
(FFT), so that the frequency spacing between the points at the FFT
output becomes .DELTA.f.sub.u=B/N.sub.u. The quantity
.DELTA.f.sub.u is referred to as the subcarrier spacing of user u.
Each user u utilizes some M.sub.u subcarriers with indices
{M.sub.u,0, . . . , M.sub.u,0+M.sub.u,0-1} out of the possible
N.sub.u for a duration of L.sub.u OFDM symbols, while the remaining
subcarriers are left empty for use by other users. Symbols that are
known by the gNB, commonly referred to as pilot symbols, are
transmitted during some subcarriers of some OFDM symbols for time
synchronization and channel estimation purposes. The pilot symbols
of user u are contained in the sparse matrix P.sub.u.di-elect
cons..sup.M.sup.u.sup.L.sup.u of which nonzero element indices are
defined as the element of a set .sub.u. The single carrier (SC)
data symbols of user u are contained in matrix D.sub.u.di-elect
cons..sup.M.sup.u.sup.L.sup.u, where the indices of non zero
elements of D.sub.u are defined as .sub.u{(m, l)|m.di-elect
cons..sub..ltoreq.M.sub.u.sup.*.andgate.(m, l).sub.u}. A cyclic
prefix of length K.sub.u samples is appended to each time domain
OFDM symbol to mitigate multipath propagation and prevent ISI. Of
these K.sub.u samples, T.sub.u.di-elect cons..sub..ltoreq.K.sub.u
are used for transmitter windowing, which results in a T.sub.u
sample cyclic suffix (CS) extension. The transmit pulse shape of
the mth subcarrier of the uth user is kept the same throughout all
L symbols, is contained in the vector t.sub.m,u.di-elect
cons..sup.(K.sup.u.sup.+N.sup.u.sup.+T.sup.u.sup.).times.1 of which
indexing is demonstrated in FIG. 7. It is worth noting that
t.sub.m,u.ident.t.sub.M.sub.u-m+1,u, for complexity purposes. The
tth sample of the baseband sample sequence to be transmitted by the
uth transmitter x.sub.u.di-elect
cons..sup.((K.sup.u.sup.+N.sup.u.sup.)L.sup.u.sup.+T.sup.u.sup.).times.1,
is obtained as:
x u [ t ] = t - 1 N u + K u t - 1 N u + K u + 1 m = 1 M u t m , u [
t - ( N u + K u ) ( l - 1 ) ] ( P u [ m , l ] + D u [ m , l ] ) e j
2 .pi. ( m + M u , 0 - 1 ) ( t - K u - 1 ) / N u , ( 8 )
##EQU00015##
[0061] for t.di-elect
cons..sub..ltoreq.(K.sub.u.sub.+N.sub.u.sub.)L.sub.u.sub.+T.sub.u,
t.sub.m,u[t]:=0, .A-inverted.t.di-elect
cons..sub.>N.sub.u.sub.+K.sub.u.sub.+T.sub.u and P.sub.u[m,
0]:=D.sub.u[m, 0]:=
P u [ m , L u + 1 ] := D u [ m , L u + 1 ] := 0 , .A-inverted. m
.di-elect cons. .ltoreq. M u * . ##EQU00016##
All transmitters then transmit their waveforms over the multiple
access multipath channel. The channel path gains are Raleigh fading
with Jakes' Doppler spectrum. The normalized complex channel gain
of the cluster that carries the samples transmitted by uth
transmitter and arrives at the gNB at the tth sample after a
propagation delay of .tau. samples is denoted by the complex
coefficient h.sub.u,.tau.,t. Then the tth received sample is
written as:
y[t]=n+.SIGMA..sub.u=1.sup.U.SIGMA..sub..tau.=0.sup.t-.DELTA..sup.t,u.su-
p.-1 {square root over
(.gamma..sub.u)}h.sub.u,.tau.,tx.sub.u[t-.DELTA..sub.t,u-.tau.],
t.di-elect cons..sup.* (9)
[0062] where x.sub.u[t]:=0, .A-inverted.t.di-elect
cons..sub.>(K.sub.+N.sub.u.sub.)L.sub.u.sub.+T.sub.u,
.A-inverted.t.di-elect cons..sub..ltoreq.U.sup.*, n.about.(0,1) is
the background additive white Gaussian noise (AWGN), .gamma..sub.u
is the signal-to-noise ratio (SNR) of uth user's received signal
and .DELTA..sub.t,u is the timing offset of uth user in number of
samples. The gNB then synchronizes to the signal of each user in
time domain by correlating the received samples with samples
generated only using P.sub.u.A-inverted.t.di-elect
cons..sub..ltoreq.U.sup.* and estimates .DELTA..sub.t,u.
[0063] The samples estimated to contain uth user's lth OFDM symbol
and its corresponding CP are denoted by vectors
y.sub.l,u.sup.OFDM.di-elect cons..sup.N.sup.u.sup..times.1 and
y.sub.i,u.sup.CP.di-elect cons..sup.K.sup.u.sup..times.1,
respectively, where
y.sub.l,u.sup.SYM[s]=y[(l-1)(N.sub.u+K.sub.u)+.DELTA..sub.t,u+K.sub.u+s],
s.di-elect cons..sub..ltoreq.N.sub.u.sup.*; and
y.sub.l,u.sup.CP[s]=y[(l-1)(N.sub.u+K.sub.u)+.DELTA..sub.t,u+K.sub.u+s],
s.di-elect cons..sub..ltoreq.K.sub.u.sup.*. The gNB uses a receiver
windowing pulse shape r.sub.m,u.di-elect
cons..sup.(N.sup.u.sup.+K.sup.u.sup.).times.1 calculated with a
receiver windowing duration of R.sub.u.di-elect
cons..sub..ltoreq.k.sub.u to receive the mth subcarrier of all OFDM
symbols transmitted by uth transmitter. Let the pulse shape parts
scaling the OFDM symbol and CP be denoted by
r.sub.m,u.sup.SYM.di-elect cons..sup.N.sup.u.sup..times.1 and
r.sub.m,u.sup.CP.di-elect cons..sup.K.sup.u.sup..times.1,
respectively, where r.sub.m,u.sup.SYM[s]=r.sub.m,u[K.sub.u+s],
s.di-elect cons..sub..ltoreq.N.sub.u.sup.* and
r.sub.m,u.sup.CP[s]=r.sub.m,u[s], s.di-elect
cons..sub..ltoreq.K.sub.u.sup.*. The SC symbol received in uth
user's lth OFDM symbol's mth subcarrier is then calculated as:
Y u [ m , l ] = F m + M u , 0 - 1 , N u ( y l , u SYM
.circle-w/dot. r m , u SYM + [ 0 N u - K u .times. 1 y l , u CP
.circle-w/dot. r m , u CP ] ) , ( 10 ) ##EQU00017##
[0064] where F.sub.m+M.sub.u,0.sub.-1,N.sub.u.di-elect
cons..sup.1.times.N.sup.u is the (m+M.sub.u,0-1)th row of the
normalized N.sub.u-point FFT matrix F.di-elect
cons..sup.N.sup.u.sup..times.N.sup.u. Channel frequency response
(CFR) coefficients at pilot tones are estimated as
H.sub.u[.sub.u]=Y.sub.u[.sub.u].0.P.sub.u[.sub.u]. Remaining CFE
coefficients in-between known coefficients are interpolated using
natural neighbors and the remainder are linearly extrapolated based
on boundary gradients. Finally, data symbols are equalized:
{circumflex over
(D)}.sub.u[.sub.u]=Y.sub.u[.sub.u].0.H.sub.u[.sub.u].
[0065] FIG. 8 illustrates the indexing of r and identification of
its parts r.sup.CP and T.sup.SYM within a demonstration of how the
receiver windowing operation of the present invention is
performed.
[0066] In the proposed method of the present invention, the mean
spectral efficiency of uth user is:
.eta. u = M u L u ( N u + K u ) L u + T u { log 2 ( 1 + 1 D u [ u ]
- D ^ u [ u ] .circle-w/dot. 2 ) } ( 11 ) M u N u + K u { log 2 ( 1
+ 1 D u [ u ] - D ^ u [ u ] .circle-w/dot. 2 ) } ( 12 )
##EQU00018##
[0067] as (N.sub.u+K.sub.u)L.sub.u>>T.sub.u, which converges
rather quickly for practical values used in all recent standards.
The proportional fair network spectral efficiency is then
.eta.=.PI..sub.u=1.sup.U.eta..sub.u. The present invention proposes
estimating and utilizing and {R.sub.1, R.sub.2, . . . , R.sub.U}
values that maximize .eta..Since the arguments of the proposal
involve modifying neither the response allocation nor the symbol
and CP durations, the focus is solely on the expectation:
.eta. ~ u = { log 2 ( 1 + 1 D u [ u ] - D ^ u [ u ] .circle-w/dot.
2 ) } , ( 13 ) ##EQU00019##
[0068] thus reducing the problem to
{ T 1 * , T 2 * , , T 1 * R 1 * , R 2 * , , R 1 * } = arg max T 1 ,
T 2 , , T U R 1 , R 2 , , R U u = 1 U .eta. ~ u ( 14 )
##EQU00020##
[0069] subject to T.sub.u, R.sub.u.di-elect
cons..sub..ltoreq.K.sub.u, .A-inverted.t.di-elect cons.{1, 2, . . .
, U}.
[0070] A review of the system model reveals that while {tilde over
(.eta.)}.sub.u depends heavily on R.sub.u, it also depends on the
transmit window durations of all users. An analytical solution to
this high complexity discrete multivariate optimization problem was
not yet shown to exist. However, the power offset across users
sharing adjacent bands can be inferred to be a useful metric upon
careful investigation of the system model in making an initial
guess for the solution of the problem.
[0071] in another exemplary embodiment, the system model has been
realized with U=2 users transmitting a frame of 140 OFDM symbols to
a base station. 256 realizations with independent and random user
data and instantaneous channels were generated and window durations
maximizing .PI..sub.u=1.sup.U{tilde over (.eta.)}.sub.u were
calculated using coordinate descent optimization for various power
offset values for each instant. Known parameters for link level
waveform evaluation under 6 GHz were used, when possible. Users
sample B=15.36 MHz with N.sub.1=2N.sub.2=1024-point FFTs, making
.DELTA.f.sub.1=15 kHz and .DELTA.f.sub.2=30 kHz. Both users utilize
a normal CP overhead of 6.7% with no additional extension for
windowing, thus conversing SG new radio (NR) frame structure. User
1 symbolizes an IoT device and experience a tapped delay line
(TDL)-C channel power delay profile (PDP) with 300 ns RMS delay
spread, and 3 km/h mobility; whereas user 2 symbolizes a slow
vehicle and experience a TDL-B channel PDP with 100 ns RMS delay
spread, and 30 km/h mobility and
.DELTA..sub.t-1-.DELTA..sub.t-2=128. P.sub.u and .sub.u are applied
from the physical uplink shared channel (PUSCH) demodulation
reference signal design without transform preceding, configuration
type 1, mapping type A, UL-DIMS-add-pos equals 1 for a PUSCH
duration of 14 symbols and single-symbol DM-RS. .gamma..sub.2=20 dB
in all instances, whereas .gamma..sub.1 is swept from 10 dB to 20
dB. The presented results are the mean of the shown metric obtained
over all realizations.
[0072] For M.sub.1=2M.sub.2=504, by utilizing such resource
allocation, either user affects and gets affected from the other
user equally in both edges of their utilized bandwidth due to the
periodicity of the EFT spectrum. This reduces the number of
variables and uncertainty in the system and allows demonstration of
the concept clearly with two users. For example, if the
recommendation of 4 resource blocks (RBs) per use was used, to
users could have only interfered with each other from only the
adjacent edges of their bandwidths, as the images of the opposite
edges are far from each other in the repeated spectra. The
demonstrate the idea in such an environment, much higher number of
users are required to cover the whole spectrum, creating further
variables in the system model and complicating it, thereby
preventing clear demonstration of the concept. Such advanced
problems and cases may be covered in future work. For example, if
the user utilizing the lower-frequency adjacent channel of a user
has a relatively lower power, whereas the user utilizing the
higher-frequency adjacent channel has a relatively higher power to
that user, the solution can be applied partially for each side of
the desired user.
[0073] The transmit and receive window durations of both users that
maximize network spectral efficiency can be seen in FIG. 9. The
values provided in FIG. 9 are presented as percentages of window
duration to the total CP duration of the corresponding user. It can
be clearly seen that the more one user outpowers the other, the
outpowering user must employ more transmit windowing and less
receiver windowing, whereas the user being outpowered must employ
less transmit windowing and more receiver windowing. The difference
of optimum durations in the no-offset case demonstrates the
dependence on the channel conditions.
[0074] FIG. 10-FIG. 12 feature several performance outcomes
comparing three windowing algorithms. The first and the baseline
algorithm is the case where both users utilize rectangular
windowing, that is T.sub.1=T.sub.2=R.sub.1=R.sub.2=0,
.A-inverted..sub..gamma..sub.1,.sub..gamma..sub.2. The second
algorithm is a simplified version, wherein regardless of the amount
of power offset between users, the outpowering user windows the
whole CP duration at the transmitter whereas the user being
outpowered windows the whole CP duration at the receiver. In the
case of no power offset, they each use half the CP duration at the
transmitter and the other half at the receiver. The third algorithm
corresponds to the optimum flexible solution that maximizes network
spectral efficiently as proposed in the present invention.
[0075] In FIG. 10, the bit-error rate (BER) performance of the
outpowering user for the three cases is shown as a function of the
power offset across users. User 2 modulates their SC data symbols
using 64-QAM. Although is has higher power than the outpowered
signal, the outpowered signal still interferes with it, of which
energy reduces as the power offset among the two signals increases.
This results in better performance and lower error rates, as shown.
The performance of the simplified algorithm compared to the
baseline algorithm shows that although the simplified algorithm
does not perform any action that explicitly improves the reception
of the current user of interest, the transmitter windowing employed
reduces the ICI outpowering user experiences due to the mobility of
the channel and improves the reception. This phenomenon is not
limited to mobility and can be observed in the presence of any time
varying effect in the communication system, such as phase noise,
which is another common problem that has a serious impact on future
high frequency devices and low cost devices. The performance of the
simplified algorithm applied in the absence of power offset is
highly similar to that of the optimum solution, implying that the
two algorithms converge in this extreme case. In any case, the
optimum solution which does not solely focus on the performance of
the outpowering user is the algorithm that improves the BER
performance of the outpowering user the most as well. It can also
be seen that the optimum solution conserves the BER performance of
the outperforming user over the outpowered user's SIR.
[0076] The BER performance of the outpowered user as a function of
their SNR can be seen in FIG. 11. User 1 always modulates their SC
data symbols using 16-QAM, regardless of instantaneous SNR. It
should be noted that except for the no-poweroffset case, actions
performed in the simplified proposed algorithm are solely focused
to maximize the performance of the outpowered user by utilizing all
resources to solely minimize the interference affecting it. It can
be seen that both the simplified and the optimum proposed
algorithms yield very similar performance that outperforms that of
the baseline receiver in any case as expected. Therefore, one can
conclude from FIG. 10 and FIG. 11 that the proposed optimum
algorithm aims to maximize the performance delivered to the
outpowering user without sacrificing from the performance of the
outpowered user as a result of the fairness implemented in the
algorithm.
[0077] The proportional fair network spectral efficiency n obtained
using each algorithm as a function of the outpowered user's SNR is
shown in FIG. 12. It can be seen that even the simplified proposed
algorithm has a gain of at least 0:43 (bit/s)/Hz gain over the
baseline algorithm, while the optimum algorithm provides further
gain. Another observation is that the simplified and optimum
algorithms differ most at low power offset values and converge to
the same values for no power offset or relatively higher power
offset values.
[0078] This exemplary embodiment demonstrates the concept of 5G
frame structure compliant power offset based extension less
windowing to maximize network spectral efficiency. The more one
user outpowers the other, the outpowering users must window
available extensions more at the transmitter side and less at the
receiver side to help reduce their impact on the network, whereas
the outpowered users must do the opposite and focus on improving
their own performance. The optimum window durations are highly
dependent on the power offset across users utilizing adjacent
channels, but also depend on the channel conditions and resource
allocation. Compared to the simplified "window whole CP on one side
depending on the sign of the power offset" solution, finding the
optimum solution allows improving the performance of the
outpowering user while conserving the performance of the outpowered
user, regardless of the amount of power offset. Depending on the
channel conditions and the severity of time varying effects in the
channel and the hardware, transmitter windowing also improves the
spectral efficiency of the user applying it. This phenomenon is
observable even for slower speeds of 30 km/h. Finding optimum
windowing durations is most beneficial when the power offset
between adjacent users is less than 6 dB; if the power offset is
more, windowing whole CP in either side yields similar results. The
performance analysis of the proposed idea for other resource
allocation scenarios by extending it to per-subcarrier transmitter
and receiver windowing durations and a solver using machine
learning techniques may be provided in the future.
[0079] The present invention may be embodied on various computing
platforms that perform actions responsive to software-based
instructions and most particularly on touchscreen portable devices.
The following provides an antecedent basis for the information
technology that may be utilized to enable the invention.
[0080] The computer readable medium described in the claims below
may be a computer readable signal medium or a computer readable
storage medium. A computer readable storage medium may be, for
example, but not limited to, an electronic, magnetic, optical,
electromagnetic, infrared, or semiconductor system, apparatus, or
device, or any suitable combination of the foregoing. More specific
examples (a non-exhaustive list) of the computer readable storage
medium would include the following: an electrical connection having
one or more wires, a portable computer diskette, a hard disk, a
random access memory (RAM), a read-only memory (ROM), an erasable
programmable read-only memory (EPROM or Flash memory), an optical
fiber, a portable compact disc read-only memory (CD-ROM), an
optical storage device, a magnetic storage device, or any suitable
combination of the foregoing. In the context of this document, a
computer readable storage medium may be any non-transitory,
tangible medium that can contain, or store a program for use by or
in connection with an instruction execution system, apparatus, or
device.
[0081] A computer readable signal medium may include a propagated
data signal with computer readable program code embodied therein,
for example, in baseband or as part of a carrier wave. Such a
propagated signal may take any of a variety of forms, including,
but not limited to, electro-magnetic, optical, or any suitable
combination thereof. A computer readable signal medium may be any
computer readable medium that is not a computer readable storage
medium and that can communicate, propagate, or transport a program
for use by or in connection with an instruction execution system,
apparatus, or device. However, as indicated above, due to circuit
statutory subject matter restrictions, claims to this invention as
a software product are those embodied in a non-transitory software
medium such as a computer hard drive, flash-RAM, optical disk or
the like.
[0082] Program code embodied on a computer readable medium may be
transmitted using any appropriate medium, including but not limited
to wireless, wire-line, optical fiber cable, radio frequency, etc.,
or any suitable combination of the foregoing. Computer program code
for carrying out operations for aspects of the present invention
may be written in any combination of one or more programming
languages, including an object oriented programming language such
as Java, C#, C++, Visual Basic or the like and conventional
procedural programming languages, such as the "C" programming
language or similar programming languages.
[0083] Aspects of the present invention are described below with
reference to flowchart illustrations and/or block diagrams of
methods, apparatus (systems) and computer program products
according to embodiments of the invention. It will be understood
that each block of the flowchart illustrations and/or block
diagrams, and combinations of blocks in the flowchart illustrations
and/or block diagrams, can be implemented by computer program
instructions. These computer program instructions may be provided
to a processor of a general purpose computer, special purpose
computer, or other programmable data processing apparatus to
produce a machine, such that the instructions, which execute via
the processor of the computer or other programmable data processing
apparatus, create means for implementing the functions/acts
specified in the flowchart and/or block diagram block or
blocks.
[0084] These computer program instructions may also be stored in a
computer readable medium that can direct a computer, other
programmable data processing apparatus, or other devices to
function in a particular manner, such that the instructions stored
in the computer readable medium produce an article of manufacture
including instructions which implement the function/act specified
in the flowchart and/or block diagram block or blocks.
[0085] The computer program instructions may also be loaded onto a
computer, other programmable data processing apparatus, or other
devices to cause a series of operational steps to be performed on
the computer, other programmable apparatus or other devices to
produce a computer implemented process such that the instructions
which execute on the computer or other programmable apparatus
provide processes for implementing the functions/acts specified in
the flowchart and/or block diagram block or blocks.
[0086] It should be noted that when referenced, an "end-user" is an
operator of the software as opposed to a developer or author who
modifies the underlying source code of the software. For security
purposes, authentication means identifying the particular user
while authorization defines what procedures and functions that user
is permitted to execute.
[0087] It will be seen that the advantages set forth above, and
those made apparent from the foregoing description, are efficiently
attained and since certain changes may be made in the above
construction without departing from the scope of the invention, it
is intended that all matters contained in the foregoing description
or shown in the accompanying drawings shall be interpreted as
illustrative and not in a limiting sense.
[0088] It is also to be understood that the following claims are
intended to cover all of the generic and specific features of the
invention herein described, and all statements of the scope of the
invention which, as a matter of language, might be said to fall
therebetween. Now that the invention has been described,
* * * * *