U.S. patent application number 13/707506 was filed with the patent office on 2016-07-21 for modal cognitive diversity for mobile communication systems.
This patent application is currently assigned to Ethertronics, Inc.. The applicant listed for this patent is ETHERTRONICS, INC.. Invention is credited to Laurent Desclos, Olivier Pajona, Sebastian Rowson, Jeffrey Shamblin.
Application Number | 20160211894 13/707506 |
Document ID | / |
Family ID | 50881431 |
Filed Date | 2016-07-21 |
United States Patent
Application |
20160211894 |
Kind Code |
A9 |
Desclos; Laurent ; et
al. |
July 21, 2016 |
MODAL COGNITIVE DIVERSITY FOR MOBILE COMMUNICATION SYSTEMS
Abstract
A system and method for antenna diversity in a communication
system are provided, the system including multiple antennas,
including at least one modal antenna, wherein each of the at least
one modal antenna has multiple modes corresponding to multiple
radiation patterns, and a processor coupled to the multiple
antennas and configured to select a mode among the multiple modes
to optimize signal quality for each time interval based on a
CQI.
Inventors: |
Desclos; Laurent; (San
Diego, CA) ; Rowson; Sebastian; (San Diego, CA)
; Shamblin; Jeffrey; (San Marcos, CA) ; Pajona;
Olivier; (Antibes, FR) |
|
Applicant: |
Name |
City |
State |
Country |
Type |
ETHERTRONICS, INC. |
San Diego |
CA |
US |
|
|
Assignee: |
Ethertronics, Inc.
San Diego
CA
|
Prior
Publication: |
|
Document Identifier |
Publication Date |
|
US 20140162566 A1 |
June 12, 2014 |
|
|
Family ID: |
50881431 |
Appl. No.: |
13/707506 |
Filed: |
December 6, 2012 |
Related U.S. Patent Documents
|
|
|
|
|
|
Application
Number |
Filing Date |
Patent Number |
|
|
13029564 |
Feb 17, 2011 |
8362962 |
|
|
13707506 |
|
|
|
|
13548895 |
Jul 13, 2012 |
8633863 |
|
|
13029564 |
|
|
|
|
12043090 |
Mar 5, 2008 |
7911402 |
|
|
13029564 |
|
|
|
|
Current U.S.
Class: |
1/1 |
Current CPC
Class: |
H04B 7/0802 20130101;
H04B 7/0871 20130101; H04B 7/0848 20130101; H04B 7/10 20130101;
H04B 7/0404 20130101 |
International
Class: |
H04B 7/04 20060101
H04B007/04 |
Claims
1. A communication system, comprising: a plurality of antennas,
comprising at least one modal antenna, wherein each of the at least
one modal antenna has a plurality of modes corresponding to a
plurality of radiation patterns; and a processor coupled to the
plurality of antennas and configured to select a mode among the
plurality of modes to optimize signal quality for each time
interval based on a channel quality indicator (CQI).
2. The communication system of claim 1, wherein during a time
interval, the processor is configured to measure an instantaneous
CQI of the selected mode, obtain a predicted CQI value of each
mode, compare the instantaneous CQI of the selected mode to the
predicted CQI values of the other modes, and if the predicted CQI
value of one of the other modes is higher than the instantaneous
CQI of the selected mode, select the one of the other modes to
optimize the signal quality for a next time interval.
3. The communication system of claim 2, wherein the processor is
configured to select the one of the other modes if the predicted
CQI value of one of the other modes is higher than the
instantaneous CQI of the selected mode by more than a predetermined
threshold.
4. The communication system of claim 2, wherein the processor is
configured to select the one of the other modes if a result of a
post filtering process for one of the other modes is higher than
the instantaneous CQI of the selected mode by more than a threshold
determined for the time interval.
5. The communication system of claim 1, further comprising: a
lookup table associated with the processor, the lookup table
including characterization data of each mode categorized by use
conditions; and a section to detect a use condition, wherein the
processor is configured to select two or more candidate modes that
are likely to optimize the signal quality under the use condition
and select a mode among the candidate modes to optimize the signal
quality for each time interval based on the CQI.
6. The communication system of claim 1, wherein the processor is
configured to obtain one or more parameters associated with motion
of the communication system, and if the one or more parameters are
in respective predetermined ranges, select a mode that optimizes
the signal quality for each time interval based on the CQI.
7. The communication system of claim 1, wherein the processor is
configured to obtain one or more first parameters associated with
motion of the communication system and to update one or more second
parameters associated with the selection based on the first
parameters.
8. The communication system of claim 4, wherein the processor is
configured to obtain one or more first parameters associated with
motion of the communication system and to update one or more second
parameters associated with the post filtering process based on the
first parameters.
9. The communication system of claim 4, wherein the processor is
configured to obtain one or more first parameters associated with
motion of the communication system and to update one or more second
parameters associated with the post filtering process and the
threshold based on the first parameters.
10. A method for implementing antenna diversity in a communication
system, comprising: providing a plurality of antennas, comprising
at least one modal antenna, wherein each of the at least one modal
antenna has a plurality of modes corresponding to a plurality of
radiation patterns; and selecting a mode among the plurality of
modes to optimize signal quality for each time interval based on a
channel quality indicator (CQI).
11. The method of claim 10, wherein the selecting during a time
interval comprises: measuring an instantaneous CQI of the selected
mode; obtaining a predicted CQI value of each mode; comparing the
instantaneous CQI of the selected mode to the predicted CQI values
of the other modes; and if the predicted CQI value of one of the
other modes is higher than the instantaneous CQI of the selected
mode, selecting the one of the other modes to optimize the signal
quality for a next time interval.
12. The method of claim 11, wherein the selecting during a time
interval comprises: if the predicted CQI value of one of the other
modes is higher than the instantaneous CQI of the selected mode by
more than a predetermined threshold, selecting the one of the other
modes to optimize the signal quality for a next time interval.
13. The method of claim 12 wherein the selecting during a time
interval further comprises: if each of the predicted CQI values of
the other modes is lower than or equal to the instantaneous CQI of
the selected mode plus the predetermined threshold, selecting the
selected mode to stay on the same mode to optimize the signal
quality for a next time interval.
14. The method of claim 10, wherein the selecting during a time
interval comprises: measuring an instantaneous CQI of the selected
mode; obtaining a metric value based on a plurality of predicted
CQI values of each mode in a post filtering process; computing a
threshold based on parameters associated with the post filtering
process; comparing the instantaneous CQI of the selected mode to
the metric values of the other modes; and if the metric value of
one of the other modes is higher than the instantaneous CQI of the
selected mode, selecting the one of the other modes to optimize the
signal quality for a next time interval.
15. The method of claim 14, wherein the selecting during a time
interval comprises: if the metric value of one of the other modes
is higher than the instantaneous CQI of the selected mode by more
than the threshold, selecting the one of the other modes to
optimize the signal quality for a next time interval.
16. The method of claim 10, further comprising: storing
characterization data of each mode categorized by use conditions in
a lookup table; detecting a use condition; and pre-selecting two or
more candidate modes that are likely to optimize the signal quality
under the use condition to select a mode among the candidate modes
to optimizes the signal quality for each time interval based on the
CQI.
17. The method of claim 10, further comprising: obtaining one or
more parameters associated with motion of the communication system,
wherein if the one or more parameters are in respective
predetermined ranges, the selecting a mode that optimizes the
signal quality for each time interval based on the CQI is carried
out.
18. The method of claim 17, wherein If at least one of the one or
more parameters is out of the predetermined range, the selecting is
stopped until the one or more parameters are in the respective
predetermined ranges.
19. The method of claim 14, further comprising: obtaining one or
more first parameters associated with motion of the communication
system, wherein if at least one of the one or more first parameters
is out of the predetermined range, the parameters associated with
the post filtering process are adjusted.
20. The method of claim 10, wherein the selecting further
comprises: storing CQI values and coefficients measured during an
initialization period; updating the stored CQI values and
coefficients during each time interval; and using the stored and
updated CQI values and coefficients to obtain a predicted CQI value
for each mode.
21. The method of claim 20, wherein the selecting further
comprises: refreshing the stored and updated CQI values and
coefficients when a refresher criterion is met.
Description
CROSS REFERENCE
[0001] This application is a CIP of U.S. Ser. No. 13/029,564, filed
Feb. 17, 2011, and titled "ANTENNA AND METHOD FOR STEERING ANTENNA
BEAM DIRECTION", which is a CON of U.S. Ser. No. 12/043,090, filed
Mar. 5, 2008, titled "ANTENNA AND METHOD FOR STEERING ANTENNA BEAM
DIRECTION", now issued as U.S. Pat. No. 7,911,402; and
[0002] a CIP of U.S. Ser. No. 13/548,895, filed Jul. 13, 2012,
titled "MODAL ADAPTIVE ANTENNA USING PILOT SIGNAL IN CDMA MOBILE
COMMUNICATION SYSTEM AND RELATED SIGNAL RECEIVING METHOD",
[0003] the contents of each of which are hereby incorporated by
reference.
BACKGROUND OF THE INVENTION
[0004] As new generations of handsets and other wireless
communication devices become smaller and embedded with increased
applications, new antenna designs, system configurations and
controlling algorithms are required to enable new capabilities and
to improve QOS.
[0005] Antenna diversity schemes are used to improve the quality
and reliability of a wireless communication link. In many
instances, the line of sight between a transmitter and a receiver
becomes blocked or shadowed with obstacles such as walls and other
objects. Each signal bounce may introduce phase shifts, time
delays, attenuations and distortions, which ultimately interfere at
the receiving antenna. Destructive interference in the wireless
link is problematic and results in degradation of device
performance. A signal quality metric is often used to assess the
quality of signals. Examples of such quality metrics include
signal-to-noise ratio (SNR), signal to interference-plus-noise
ratio (SINR), receive signal strength indicator (RSSI), bit error
rate (BER) and other metrics, which are called channel quality
indicators (CQI). An antenna diversity scheme can mitigate
interference from multipath environments by monitoring one or more
CQIs. Antenna diversity can be implemented generally in several
forms, including spatial diversity, pattern diversity and
polarization diversity, for example.
[0006] Spatial diversity for reception includes multiple antennas
having similar characteristics, which are physically spaced apart
from one another. In multipath propagation conditions, as
encountered with a blocked or shadowed line of sight path, each of
the multiple receive antennas experiences a different fading
characteristic. Accordingly, where a first antenna experiences a
significant reduction in signal reception, the second antenna is
likely to receive an effective signal. Collectively, the spatial
diversity scheme can provide a robust link. Spatial diversity for
transmission is also effective, although link improvements may be
needed in certain cases for the receive side of the base
station.
[0007] Pattern diversity generally includes two or more co-located
antennas with distinct radiation patterns. This technique utilizes
antennas that generate directive beams and are usually separated by
a short distance. Collectively, these co-located antennas are
capable of discriminating a large portion of angle space and may
additionally provide relatively higher gain compared to an
omnidirectional antenna.
[0008] Polarization diversity generally includes paired antennas
with orthogonal polarizations. Reflected signals can undergo
polarization changes depending on the medium through which they are
traveling. By pairing two complimentary polarizations, this scheme
can immunize a system from polarization mismatches that would
otherwise cause signal fade.
[0009] Each of the above diversity schemes requires one or more
processing techniques to effectuate antenna diversity, such as
switching, selecting and combining. Switching is one of the simple
and efficient processing techniques and generally includes
receiving a signal from a first antenna until the signal level
fades below a threshold level, at which point active components
such as switches engages the second antenna for communication with
the receiver. Selecting is a processing technique that determines
an optimal signal for utilization by the receiver during each
predetermined time interval. Both selecting and switching
techniques may utilize active components, such as switches, to
select the optimal signal based on one or more CQIs. The selecting
and switching techniques may be collectively called a switching
technique wherein the selection of the signal for utilization is
carried out by controlling the switches or other active components
coupled to the antennas.
[0010] Combining is a processing technique wherein the multiple
signals are weighted and combined into a signal for communication
with the receiver. The weights are adjusted such that a main lobe
of a receiving beam pattern is toward the desired direction and a
null of the receiving beam pattern is toward the interference
direction. Examples of combining techniques include a minimum mean
squared error (MMSE) combining technique, a maximum ratio combining
(MRC) technique and an equal gain combining (EGC) technique. An
exemplary algorithm to carry out each of these combining techniques
may be briefly summarized as follows. In the MMSE technique, the
signals in paths are weighted where the weights are chosen to
provide a minimum mean square error between the combined voltage
stream and the signal. In the MRC technique, the signals in paths
are weighted where the weights are chosen to be proportional to the
respective signal amplitudes to maximize the output SNR. The
weighted signals are then multiplied by respective phase factors
prior to summing so that the signals are added in phase to maximize
the gain. The EGC technique is a simplified version of the MRC
technique, wherein the signals are weighted with the same factor
and then multiplied by the phase factors.
[0011] Although the above techniques have been described for
reception, their analogs are possible for transmit functions.
Receive (Rx) diversity refers to configurations where a diversity
scheme is applied for signal reception; and transmit (Tx) diversity
refers to configurations where a diversity scheme is applied for
signal transmission
BRIEF DESCRIPTION OF THE DRAWINGS
[0012] FIG. 1 illustrates an example of a modal antenna.
[0013] FIG. 2(a) illustrates a radiation pattern associated with
the modal antenna of FIG. 1 in the first state.
[0014] FIG. 2(b) illustrates a radiation pattern associated with
the modal antenna of FIG. 1 in the second state.
[0015] FIG. 3 illustrates an example of a communication system
including multiple antennas, wherein at least one of the multiple
antennas is a modal antenna and the others are conventional
non-modal antennas; or all the multiple antennas are modal
antennas.
[0016] FIG. 4 illustrates an example of a hardware configuration of
an antenna system having one modal antenna and one conventional
non-modal antenna.
[0017] FIG. 5 illustrates an example of a communication system
including a conventional non-modal antenna and a modal antenna.
[0018] FIG. 6 illustrates an example of the lookup table (LUT) for
the case a four-mode modal antenna.
[0019] FIG. 7 illustrates an example of the modal cognitive
diversity (MCD) procedure using the CQI prediction algorithm for a
case of having four modes with the filter length of three.
[0020] FIG. 8 is a flow chart illustrating a first example of the
MCD process.
[0021] FIG. 9 is a flow chart illustrating a second example of the
MCD process.
[0022] FIG. 10 is a flow chart illustrating a third example of the
MCD process.
DETAILED DESCRIPTION OF THE PREFERRED EMBODIMENTS
[0023] In Rx diversity for a conventional wireless mobile device, a
main antenna is typically used to cover the transmit and receive
bands. One or more diversity antennas are added in the device to
cover the receive band and to counteract the effect of propagation
and multipath fading, thus requiring additional receive chains. The
diversity functions may be effectuated by means of the switching
technique described earlier to switch the receive path from one
antenna to the other to improve the signal quality during a given
time interval according to a decision based on SNR or other CQI
parameters. Another scheme involves the combining technique
described earlier to constructively combine signals for enhancing
overall signal quality based on SNR or other CQI parameters. In a
conventional spatial or pattern diversity architecture with
multiple antennas, size constraints may become significant and
coupling among the antennas as well as between the antennas and
nearby electronics of a communication device may deteriorate
transmission and reception qualities. Additionally, efficiency may
deteriorate in many instances where multiple paths are energized
and power consumption increases. Implementing two, three or more
diversity antennas may provide system flexibility, but the volume
and area requirements become increasingly severe as additional
antennas, associated components and transmission lines are
needed.
[0024] A modal antenna, also referred to as a null-steering
antenna, can generate multiple modes corresponding to multiple
radiation patterns, equivalent to having a set of antennas with the
same feed. By using the modal antenna capable of generating
different radiation patterns, it is possible to exploit a priori
knowledge of antenna nulls and lobes in the different modes for
steering the beam to have nulls in dominant interference directions
while keeping gain in desired directions. Accordingly,
implementation of a modal antenna may allow for pattern and
polarization diversity with one antenna, thereby requiring less
volume and area than using multiple antennas for diversity.
Further, the use of a modal antenna in Rx diversity requires less
number of receiver chains than using multiple antennas. This
document describes antenna diversity using one or more modal
antennas and a method to dynamically find an optimal mode of the
modal antenna for signal processing based on predicted CQI values.
Examples of structures and implementations of the modal antennas
are provided in U.S. Pat. No. 7,911,402, entitled "ANTENNA AND
METHOD FOR STEERING ANTENNA BEAM DIRECTION," issued on Mar. 22,
2011. The contents of the above patent are incorporated herein by
reference and are summarized as follows.
[0025] FIG. 1 illustrates an example of a modal antenna 100, which
includes an Isolated Magnetic Dipole.TM. (IMD) element 104 placed
on a ground plane 108, a first parasitic element 112 coupled to an
first active element 113, and a second parasitic element 116
coupled to a second active element 117. The active elements 113 and
117 may include switches that either electrically connect (short)
or disconnect (open) the parasitic elements 112 and 116 to the
ground plane 108. This structure allows for two different modes of
operation with a common frequency corresponding to a first state
where the parasitic elements 112 and 116 are shorted to the ground
and a second state where the parasitic elements 112 and 116 are
open. FIG. 2(a) illustrates a radiation pattern 204 associated with
the antenna 100 in the first state; and FIG. 2(b) illustrates a
radiation pattern 208 in the second state, which shows a
ninety-degree shift in direction as compared to the radiation
pattern 204. Thus, by controlling the active elements 113 and 117
of the modal antenna 100, the operation of two modes can be
obtained at the same frequency. The control scheme can be extended
for three or more multi-mode operations by incorporating, for
example, tunable elements in the active elements for variable
control and additional active elements for matching. Examples of
these active elements include switches, tunable capacitors, tunable
phase shifters, diodes, micro-electro-mechanical system (MEMS)
switches, MEMS tunable capacitors, and transistors including a
metal oxide semiconductor field effect transistor (MOSFET), a metal
semiconductor field effect transistor (MESFET), a pseudomorphic
high electron mobility transistor (pHEMT), a heterojunction bipolar
transistor (HBT) or of other suitable technologies.
[0026] FIG. 3 illustrates an example of a communication system 300
including multiple antennas 304, 305 . . . and 306, a processor 312
and a radio frequency (RF) circuit 316, wherein at least one of the
multiple antennas 304, 305 . . . and 306 is a modal antenna and the
others are conventional non-modal antennas; or all the multiple
antennas 304, 305 . . . and 306 are modal antennas. The antenna
configuration in FIG. 3 may be used for spatial, pattern and/or
polarization diversity. Rx diversity can be implemented by
configuring two or more of the multiple antennas 304, 305 . . . and
306 for reception; Tx diversity can be implemented by configuring
two or more of the multiple antennas 304, 305 . . . and 306 for
transmission. Tx and Rx diversities can be collocated in the
system, by configuring two or more antennas for transmission and
other two or more antennas for reception. The multiple antennas
304, 305 . . . and 306 are coupled to the processor 312 through the
RF circuit 316. The RF circuit includes various modules and
components for processing signals, such as a transceiver, filters,
duplexers, switches, power amplifiers, matching networks and so on.
The processor 312 may be included in a baseband, a CPU or an
application CPU, or associated with the transceiver or the
antennas. The processor 312 may be configured to store data,
specifications and other prerequisites as well as to perform
algorithms to control the antennas and other components and modules
for signal processing.
[0027] FIG. 4 illustrates an example of a hardware configuration of
an antenna system having one modal antenna 404 and one conventional
non-modal antenna 408. The modal antenna 404 is located at one edge
of the board, having an IMD 405, a parasitic element 406 and an
active element 407 coupled to the parasitic element 406. The
non-modal antenna 408 is located at the other edge of the board,
having a main radiator 409. In this example, the modal antenna 404
is coupled to a transceiver 1 412, and the non-modal antenna 408 is
coupled to a transceiver 2 416, for processing Tx or Rx signals. A
processor 420 is coupled to the transceiver 1 412 and the
transceiver 2 416 for controlling signals and performing algorithms
for diversity and other functions. Furthermore, in this example,
the processor 420 is directly coupled to the active element 407 to
configure the parasitic element 407 to generate multiple modes
associated with the modal antenna 404 by the switching control, and
the control signal path 424 is indicated by the dashed line in this
figure.
[0028] FIG. 5 illustrates an example of a communication system 500
including a conventional non-modal antenna 504, a modal antenna
506, a processor 512 and a radio frequency (RF) circuit 516. This
configuration represents a simple case of the multi-antenna system
shown in FIG. 3, having only two antennas, one of which is a modal
antenna. A look-up table (LUT) 520 may be added to provide input
information for the processor 512. The LUT 520 may contain measured
data and be stored in a memory in the processor 512 or in a
different location coupled to the processor 512. The antenna
configuration in FIG. 5 may be used for spatial, pattern and/or
polarization diversity. Either Rx diversity or Tx diversity may be
implemented by configuring two antennas 504 and 506 for reception
or transmission. The antennas 504 and 506 are coupled to the
processor 512 through the RF circuit 516. The RF circuit 516
includes various modules and components for processing signals,
such as a transceiver, filters, duplexers, switches, power
amplifiers, matching networks and so on. The processor 512 may be
included in a baseband, a CPU or an application CPU, or associated
with the transceiver or the antennas. The processor 512 may be
configured to store data, specifications and other prerequisites as
well as to perform algorithms to control the antennas and other
components and modules for signal processing. In the example in
FIG. 5, the modal antenna 506 is controlled by the processor, as in
the hardware configuration example of FIG. 4, the processor
controlling the on/off of the active elements to configure the
parasitic elements to generate multiple modes corresponding to
multiple radiation patterns. Such a modal antenna tuning means may
be included in the processor 512 as in FIG. 5, coupled with the
modal antenna 506 to form a "plug-and-play" modal antenna system,
or separated as a modal antenna tuning module.
[0029] The LUT 520 may include characterization data of the modal
antenna 506, such as total radiated power (TRP), total isotropic
sensitivity (TIS), radiation patterns and so on, which can be
measured in advance for various use conditions. The radiation
pattern may be stored with the (.rho.,.theta.) coordinate having N
points (300-500 points, for example). The use conditions may
include positioning of a head and/or a hand of a user including the
possible distances and angles with respect to the handset. FIG. 6
illustrates an example of the LUT 520 for the case a four-mode
modal antenna. Under each use condition, each mode is listed along
with the characterization data.
[0030] During operation, the use condition may be detected by one
or more sensors such as a proximity sensor, a motion sensor, a
light sensor, a pressure sensor or other types of sensors.
Information on the use condition may be inputted to the LUT 520 as
indicated by dashed line 524. Based on the information and the
corresponding characterization data of each mode of the modal
antenna 506, two or more modes that are likely to provide optimal
signal properties can be chosen under the use condition. The use
condition varies with time, and thus the processor 512 may be
configured to perform the mode-selection procedure dynamically
based on a predetermined time interval and/or a change
threshold.
[0031] Referring back to FIG. 3, where the system includes at least
one modal antenna, antenna diversity can be achieved by selecting
one of the antennas or one mode out of the multiple modes that the
modal antenna provides. Alternatively, the combining technique can
be utilized to constructively combine multiple signals from
respective chains to enhance the overall signal strength and
quality for diversity. In either case, it is necessary first to
find a mode that provides the optimal signal quality among the
multiple modes of the modal antenna. Furthermore, referring back to
FIG. 5, where based on the LUT 520 the processor 512 selects two or
more candidate modes that are likely to provide optimal signal
properties, it is still necessary to find a mode that provides the
optimal signal quality among the candidate modes during operation.
The procedure to select such an optimal mode needs to be carried
out dynamically since the use conditions, movement of the handset,
incoming signal properties, base station conditions and various
other conditions and surroundings vary with time. In the following,
a method to dynamically find an optimal mode among multiple modes
of a modal antenna for diversity processing is described based on a
technique involving CQI prediction. The method is thus termed modal
cognitive diversity (MCD) in this document.
[0032] An example of a CQI prediction algorithm is described by
Touheed et al., "Predictive CQI Reporting for HSDPA," IEEE 19th
International Symposium, 15-18 Sep. 2008, which is briefly
summarized below. A linear predictor based on a
normalized-least-mean-square (NLMS) adaptive filter can be used to
derive future CQI estimates. In this algorithm, the vector W(n)
represents the filter tap weight vector (coefficients of filter)
and the vector X(n) represents the filter input vector (CQI values)
at time n as follows:
W(n)=[W.sub.1(n), W.sub.2(n), . . . , W.sub.M(n)].sup.T
X(n)=[CQI(n-1), CQI(n-2), . . . , CQI(n-M)].sup.T.
Here (.).sup.T represents transpose operator, and M represents the
filter length. The NLMS adaptation equation gives the coefficients
of filter at time n+1 as follows:
W ( n + 1 ) = W ( n ) + .mu. X ( n ) e ( n ) X ( n ) X H ( n ) . Eq
. ( 1 ) ##EQU00001##
Here .mu. is the step size parameter, e(n) is the CQI prediction
error, and (.).sup.H represents Hermitian transpose. The future
value of CQI, CQI(n+1), can be predicted at time n as:
CQI(n+1)=X(n+1).W.sup.H(n+1), Eq. (2)
where
X(n+1)=[CQI(n), CQI(n-1), . . . , CQI(n+1-M)].sup.T. Eq. (3)
Thus, the above equation Eq. (2) for CQI(n+1) indicates that the
CQI value at time n+1 can be predicted based on the coefficients of
filter W(n) that depends on the quantities during the filter length
M up until time n as shown in Eq. (1) and the CQI values during the
filter length M up until time n as shown in Eq. (3).
[0033] Pilot signals can be used to measure and calculate CQIs in
the CQI prediction algorithm. In cellular and personal
communication services (PCS) systems based on code division
multiple access (CDMA), for example, a pilot signal is used for
synchronization, coherent detection, soft handoff, maintaining
orthogonally between base stations and so on. A pilot signal is a
carrier modulated by a high chip rate pulse-shaped pseudo noise
(PN) sequence and is transmitted on the forward link (from a base
station to a mobile unit). All mobile units have the prior
knowledge of the periodic PN sequence.
[0034] In the present MCD scheme, the CQI prediction algorithm is
carried out for each mode of a modal antenna at each time step to
determine the expected (predicted) CQI for each mode j, CQIj(n+1).
Specifically, the filter input vector X(n) contains CQI values
which are either previous measurements CQIi(n-k) (when the mode j
corresponding to the filter j was selected), or previous predicted
CQI values for this mode CQIj(n-k) (when the mode j corresponding
to the filter j is currently selected). It can also contain a
combination of CQIi(n-k) and CQIj(n-k) during transition phases
which occur when the decision is made to switch from one mode to
another mode.
[0035] At a given time step n, the instantaneous CQI for the
currently selected i-th mode, CQIi(n), is measured and compared to
the predicted CQI for the j-th mode, CQIj(n+1), and the mode
switches to the j-th mode if:
CQI.sub.i(n)<CQI.sub.j(n+1)-.DELTA., Eq. (4a)
where .DELTA. is a predetermined threshold and i.noteq.j. The
threshold may be predetermined such that the switch from one mode
to another gives a meaningful, good enough CQI improvement. If the
expected CQI improvement is less than the threshold, or the other
modes provide the predicted CQI values lower than the instantaneous
CQI of the i-th mode, the mode stays the same.
[0036] The above algorithm can be extended to predict CQI values in
the several future steps to be used for the decision making as
expressed in the following:
CQI.sub.i(n)<f(CQI.sub.j(n+1), . . . , CQIj(n+k))-.DELTA. Eq.
(4b)
CQI.sub.i(n)<f(CQI.sub.j(n+1), . . . ,
CQIj(n+k))-.DELTA..sub.j(f(j)), Eq. (4c)
where k represents the number of future steps, and f(.) is a
function of the predicted CQI values in the future k steps. The
function, f, may represent an operation of taking an average,
selecting the best or worst predicted CQI value among the k-number
predicted CQI values, obtaining an extrapolated CQI value, or other
suitable operations. The above extended process may be referred to
as post filtering process.
[0037] Specifically, Eq. (4a) is extended as Eq. (4b) so that the
measured CQIi(n) is not only compared to CQIj(n+1) as in Eq. (4a),
but to the result of taking into account several predictions
forward, (CQIj(n+1) . . . CQIj(n+k)). Eq. (4b) is modified as Eq.
(4c), where the .DELTA. threshold value is no longer fixed but can
vary over time by taking into account various parameters for each
filter corresponding to a mode, for example, the result of the
previous mode switch, the time since the coefficients of the filter
have been updated, or other dynamically varying properties.
[0038] The above algorithm can be further extended so that the
filter input vector X(n) gets conditioned before entering the
filtering process, through additional steps, to modify its range or
scale in order to optimize the predictive filtering process.
Typically the DC or the magnitude offset contained in the original
signal can be subtracted from the signal before entering into the
filtering process. These additional steps may be referred to as
"pre filtering process." The DC or the magnitude offset may be
added to the filter's output signal during the post processing
process.
[0039] FIG. 7 illustrates an example of the MCD procedure using the
CQI prediction algorithm for a case of having four modes with the
filter length of three. This example involves the prediction of one
step forward as represented by Eq. (4a), and is explained here for
illustrative purposes. The tick marks in the time line represent
arrivals of pilot signals. The first 12 pilot signals are used to
measure CQI values and corresponding coefficients of filter, which
are stored for initialization. Fictitious numbers representing CQI
values are inputted for mode 1, mode 2, mode 3 and mode 4 for
illustrative purposes in this figure. At time t0, the CQI
prediction algorithm calculates the predicted CQI, CQI, for each
mode based on the CQI values and coefficients measured and stored
during the initialization period. The mode having the best
predicted CQI is selected. In this example, mode 4 has the best
predicted CQI of 6. At time t1, the instantaneous CQI is measured
for mode 4, giving the value of 7 in this example. The measured CQI
and corresponding coefficients are stored in the mode 4 memory
section. The CQI prediction algorithm calculates the next predicted
CQI, CQI, for each mode based on the CQI values and coefficients
during the last three time segments. The switching threshold,
.DELTA., in Eq. (4) is predetermined to be 1.9 in this case. The
instantaneous CQI for mode 4, which is measured to be 7, is
compared against each of the predicted CQI values for the four
modes. Since there is no mode giving a predicted CQI value that
exceeds the instantaneous CQI of mode 4 by more than the threshold
.DELTA.=1.9, the selected mode remains the same, which is mode 4.
At time t2, the instantaneous CQI is measured for mode 4, giving
the value of 6 in this example. The measured CQI and corresponding
coefficients are stored in the mode 4 memory section. The CQI
prediction algorithm calculates the next predicted CQI, CQI, for
each mode based on the CQI values and coefficients during the last
three time segments. The result shows that mode 2 gives the best
predicted CQI of 8, which exceeds the instantaneous CQI of mode 4,
which is 6, by more than the threshold .DELTA.=1.9. Thus, mode 2 is
selected. At time t3, the instantaneous CQI is measured for mode 2,
giving the value of 5 in this example. The measured CQI and
corresponding coefficients are stored in the mode 2 memory section.
The CQI prediction algorithm calculates the next predicted CQI,
CQI, for each mode based on the CQI values and coefficients during
the last three time segments. The result shows that mode 1 gives
the best predicted CQI of 9, which exceeds the instantaneous CQI of
mode 2, which is 5, by more than the threshold .DELTA.=1.9. Thus,
mode 1 is selected. The similar procedure is carried out at the
subsequent time steps, predictively determining the mode that gives
the best CQI during each time interval.
[0040] As explained in the above example, the memory section of
each mode includes CQI values during the filter length, and
continuously updated with measured or predicted values. Over a
certain time period, the stored information may become obsolete,
not reflecting the actual CQI variation for the mode, due, for
example, to lack of convergence or stability of the algorithm.
Accordingly, the memory may need to be refreshed by repeating the
initialization procedure to measure and store new CQI values and
coefficients. The filter length may be of the original length for
the entire refresher, or less for a partial refresher. The
refresher process may be configured to start when a certain
predetermined criterion is met, for example, when the predicted CQI
value becomes less than a certain predetermined value.
[0041] The training (initialization or refresher) periods for
different modes can be adjusted to follow different schemes. The
period may be fixed consecutive steps for each mode (as described
in the above example of FIG. 7), or may be divided into several
smaller periods which are alternated between different modes. For
example, instead of 3 steps for mode 1 and then 3 steps for mode 2,
it could be 2 steps for mode 1, 2 steps for mode 2, and then 1 step
for mode 1 and 1 step for mode 2. Alternatively, the training
period for a mode may be determined by a condition on the error
function e(n) in Eq. (1), such that the training on each mode will
last until a predetermined error threshold is reached.
[0042] The step size parameter, .mu., in Eq. (1) is related to
convergence of the algorithm, and may be adjusted to vary over time
for each filter corresponding to a mode in order to reduce the
training period and/or to optimize the prediction, for example. The
parameter .mu. can be adjusted according to motion of the device
(speed, acceleration, angle change, etc.) as well as use conditions
(presence of a head, a hand, etc.) and multi-path environments
(reflection, fading, shadowing, interference, etc.) to take into
account the speed of variation of the channel relative to the
device.
[0043] For real time constraints reasons, the algorithm can be
implemented in such way that the different filters and their post
filtering processes are independent of each other and run in
parallel. In such a case, the outputs of the different filters and
their post filtering processes are configured to synchronize at the
decision stage level. The independent filters can have different
parameters, such as different filter lengths M, step size
parameters .mu., numbers of future prediction steps k, thresholds
.DELTA., and so on.
[0044] There are instances when the handset is experiencing unusual
local movements, such as shaking, fast speed or a sudden change in
motion or speed. In another instances, the handset may be sitting
idle on a table surface. Sensors such as a proximity sensor, a
motion sensor, a light sensor or a pressure may be used to detect
speed, acceleration, angle change and other motion parameters.
Global positioning system (GPS) data may also be used to determine
the speed, acceleration, and direction of the mobile device. These
motion parameters associated with the device movements may be
detected and used to interrupt the MCD process to avoid unnecessary
mode selection procedures. This is because the CQI may be changing
so rapidly that it is meaningless to select the best modes for
short time intervals individually, or not changing at all so that a
new mode selection is not necessary. The MCD process may be resumed
when the detected motion parameter falls within a certain nominal
range, for example. Alternatively, without interrupting the MCD
process, the various parameters such as the filter length M, the
step size parameter .mu., the number of future prediction steps k
in the post filtering process, and the threshold .DELTA. may be
dynamically adjusted to optimize the process depending on
information from a CPU, ASICs, sensors or other condition- or
environment-detecting means.
[0045] FIGS. 8, 9 and 10 are flow charts, each illustrating an
example of the MCD process to dynamically select the optimal mode
of a modal antenna for each time interval. FIG. 8 illustrates a
first example of the MCD process, which is process A between the
start and end steps. In step 804, CQI values and corresponding
coefficients of filter for each mode are measured using the pilot
signals and stored in the memory for initialization. These values
for each mode are measured for the filter length period. An initial
current mode is selected among the modes. Subsequent pilot signals
are used for measurement throughout the process. In step 808, the
predicted CQI, CQI, for each mode is obtained using the CQI
prediction algorithm based on the stored CQI values and
coefficients during the last filter length. In step 812, it is
determined if refreshing of the memory is needed based on a certain
refresher criterion. If needed, the process goes back to step 804
to measure and store new CQI values and corresponding coefficients
for each mode. The filter length for the measurement may be of the
original length for the entire refresher, or less for a partial
refresher. The refresher process may be optional and may be skipped
if the original initial values provide overall convergence and
stability with the algorithm. If the refresher is not needed, the
process proceeds to step 816 to measure the instantaneous CQI of
the currently selected mode. In step 820, if any other mode has the
predicted CQI higher than the instantaneous CQI of the current
mode, the current mode is switched to the mode that has the higher
predicted CQI. A threshold for the CQI comparison may be
predetermined such that the switch from one mode to another gives a
meaningful, good enough CQI improvement. If the expected CQI
improvement is less than the threshold, or the other modes provide
the predicted CQI values lower than the instantaneous CQI of the
current mode, the mode stays the same. The process of selecting the
optimal mode based on the comparison between the instantaneous CQI
of the current mode and the predicted CQI values of the other modes
is repeated until the MCD process is determined to end at step 824
based on user inputs or control information.
[0046] FIG. 9 illustrates a second example of the MCD process. This
example utilizes a LUT. In step 904, the use condition is
determined based on information detected by one or more sensors,
for example. The use conditions include positioning of a head
and/or a hand of a user such as possible distances and angles with
respect to the handset. Based on the use condition, candidate modes
are selected according to the characterization data stored in the
LUT. The MCD process such as the process A of FIG. 8 follows as in
step 912 to select the optimal mode among the candidate modes. It
is checked at certain times if the use condition has changed in
step 916. If changed, the process goes back to step 904 to
determine the new use condition. New candidate modes are selected
in step 908 under the new use condition, and the process A is
repeated to select the optimal mode among the new candidate modes
in step 912.
[0047] FIG. 10 illustrates a third example of the MCD process. This
process avoids unnecessary mode selection procedures when the
device is experiencing unusual movements, such as shaking, fast
speed, sudden change in speed, sitting idle and so on. These are
the situations where the CQI may be changing so rapidly that it is
meaningless to select the best modes for short time intervals
individually, or not changing at all so that a new mode selection
is not necessary. In step 1004, motion parameters are obtained to
detect the motion of the device, such as speed, acceleration,
direction, angle change and the like. Sensors in the device and/or
the GPS system may be used for the motion detection. If the motion
parameters indicate that the motion of the device is in a
predetermined nominal range at step 1008, the process A follows in
step 1012 to select the optimal mode for each time interval. If at
least one of the motion parameters, such as speed, is out of the
nominal range, the process A is not initiated. The process may be
configured so that even during the process A, step 1004 may be
taken to obtain motion parameters with certain time intervals to
check if any change in motion has occurred. The process A continues
when the motion parameters fall in the respective nominal ranges.
If at least one of the motion parameters is out of the nominal
range, the process A is stopped until all the motion parameters get
back to respective nominal ranges.
[0048] Three examples of the MCD process are explained in the
above. Variation examples include the combination of the second and
the third examples to utilize the LUT for selecting candidate modes
for time saving as well as to detect motion parameters to avoid
unnecessary selection procedures when the device is experiencing
unusual movements.
[0049] While this document contains many specifics, these should
not be construed as limitations on the scope of an invention or of
what may be claimed, but rather as descriptions of features
specific to particular embodiments of the invention. Certain
features that are described in this document in the context of
separate embodiments can also be implemented in combination in a
single embodiment. Conversely, various features that are described
in the context of a single embodiment can also be implemented in
multiple embodiments separately or in any suitable subcombination.
Moreover, although features may be described above as acting in
certain combinations and even initially claimed as such, one or
more features from a claimed combination can in some cases be
exercised from the combination, and the claimed combination may be
directed to a subcombination or a variation of a
subcombination.
* * * * *