Note : Les descriptions sont présentées dans la langue officielle dans laquelle elles ont été soumises.
CA 02219096 1997-10-24
Diversity Antenna Selection
Field of the Invention
The present invention is directed generally to wireless
telecommunications, and in particular, it is directed to the technology of
antenna diversity, with particular relevance to diversity reception of
digital signals used in fixed wireless access applications.
Background of the Invention
In wireless telecommunications, multi-path fading is a known
cause of fluctuation in received signal levels and therefore a known cause
of deterioration in communication.
Diversity reception has been widely performed, as a method of
reducing this fading. For example if two antennas are separated at
predetermined distances from each other, there is a significantly reduced
probability that the signals from both antennas are simultaneously
attenuated to the same extent, hence achieving a higher reliability. For
example, signal to Inle~reLellce margins can be increased by as much as 10
dB when two antennas are used rather than one. Various kinds of
20 diversity reception methods are known.
In most cases, if "perfect" diversity is assumed, a system
requires apriori knowledge of the received signal strength indication (RSSI
hereafter) at all antennas which can be selected. For many prior art
diversity systems, such information can only be achieved through the use
of dual receivers. In such an arrangement, the RSSI of the signal received
on each antenna is continuously monitored, and the best antenna is
selected. This implementation adds considerable cost to a radio receiver.
Some prior art solutions teach diversity receivers which use
two or more antennas and one receiver circuit. An antenna selection
3 0 circuit switches between antennas in response to received signal strength
indication generated by the receiver. However, without a dual receiver,
CA 02219096 1997-10-24
the unused antenna must be sampled periodically, resulting in bit hits
(lost or corrupted bits) or frame hits (lost or corrupted frames) in the
received data whenever the "tested" antenna has a poor RSSI.
Furthermore, these prior art diversity systems have in the past
been of particular relevance to base stations and have been of less
relevance to user terminals in part due to the requirements of having the
antennas physically separated by a minimum distance of half a
wavelength. This physical separation of the antennas has typically made it
impractical to incorporate such techniques in small, compact mobile
10 terminals. Also, as stated, diversity techniques typically utilize dual
receivers which are acceptable for base stations but are not cost effective for
terminals.
However, in many ways the problem of multi-path fading
presents more of a problem to a fixed wireless access terminal than to a
mobile terminal because the subscriber terminal is stationary and therefore
the terminal is unlikely to be moved by the user in response to poor
reception as a result of a deep fade.
Furthermore, even in situations where multi-path fading does
not represent a significant problem, the quality of reception can still be
2 0 poor as a result of co-channel interference (CCI) in the forward link. This
problem is particularly serious in high capacity cellular networks where
the desire to reuse frequencies many times can degrade the signal to
interference ratio (SIR). However, conventional prior art diversity
solutions designed to switch antennas based on a comparison of the
received signal strength indication (RSSI) between antennas will not
typically solve such a problem.
Another problem not solved adequately by the prior art as
known by the inventor herein, is that some users may have their radio
link performance limited by the forward link signal strength while other
3 o users may have their forward link quality limited by co-channel
interference (CCI) due to a high subscriber density. Furthermore, typically,
CA 02219096 1997-10-24
there exists insufficient apriori knowledge to predict which of these two
problems is the dominant cause of poor reception. Furthermore, both of
the above impairments (i.e., fading and CCI) can occur at various times,
and independently of each other. Therefore, there exists a need for an
improved diversity system which can improve reception regardless of
which impairment causes degraded reception.
Summary of the Invention
One aspect of the invention provides a diversity selection
process, and also a terminal for carrying out the process, while also
providing for mobile assisted hand-off (MAHO) measurements. The
inventor has identified this as being particularly advantageous for fixed
wireless access (FWA) terminals, which do not move in and out of cells
and therefore have not previously been set up for hand off. By now
providing a FWA terminal with both diversity and MAHO a terminal can
improve service by switching antennas for relatively short term problems,
or the terminal allows for switching to another basestation (if available)
for persistent problems.
Another aspect of the present invention is directed to a process
and apparatus for performing antenna diversity which first evaluates the
nature of an impairment causing poor reception and then switches
antennas accordingly. Such a system is of particular suitability for fixed
wireless access terminals using digital radio communications.
It should be noted that these techniques are applicable to any
type of antenna scheme where the signals can be suitably decorrelated,
regardless of whether such decorrelation is achieved through space or
polarization diversity.
Brief Description of the Drawings
3 0 The present invention, together with further objects and
advantages thereof will be further understood from the following
CA 02219096 1997-10-24
description of example embodiments with reference to the drawings in
which:
Fig. 1 is a schematic block diagram of a wireless access terminal
incorporating a preferred embodiment of the present invention.
Fig. 2 is a flow diagram illustrating the diversity selection
process steps carried out by the baseband microcontroller of Fig. 1,
according to one embodiment of the invention.
Fig. 3 is a timing reference chart illustrating further aspects of
the process of Fig. 2 according to one embodiment of the invention.
Fig. 4 is a series of flow charts illustrating the diversity selection
process steps carried out by the baseband microcontroller according to
another embodiment'of the invention.
Fig. 5 is a flow chart illustrating the conceptual steps according
to one embodiment of the invention, for which Fig. 4 represents a specific
2 o implementation.
Detailed Description of the Preferred Embodiment
The preferred embodiment of the present invention will be
described with respect to its application within a subscriber unit such as the
fixed wireless access terminal shown in Fig. 1. The preferred embodiments
will also be described for IS54-B, TDMA-3 It should be appreciated by a
person skilled in the art that this example is used for the purposes of
illustration and the invention is also applicable to other systems.
In Fig. 1 the terminal (also called subscriber unit) comprises a
3 o radio block 10, a baseband block 60 and a voice frequency block 110. There
are two interfaces between the baseband block 60, and the voice frequency
CA 02219096 1997-10-24
block 110. The first interface called the PCM interface 175 comprises the
digitized voice frequency pulse code modulation (PCM) transmit and
receive signals, while the second interface is a bi-directional serial
communications interface 178. The radio block 10, and the baseband block
60, provide the conversion between the radio frequency and digitized
voice frequency signals. The baseband block 60 is also responsible for
handling the protocols associated with the RF link under the control of the
voice frequency block 110.
The voice frequency block 110 includes a primary user interface
which includes a display 120, a keypad 130, an alerter 150 which produces
an audio alert (e.g., ringing) and an indicator which provides visual alerts
(e.g., a light indicator indicating, for example, that an extension is off-hook
or that an incoming call has been received), and a primary handset 140.
The voice frequency block 110 also includes a secondary user interface
including RJ-11 jack 230 which acts as an extension jack for a standard
analog telephony devices. Note that an additional data jack can be
supported.
Not shown is a suitable DC power source. This can comprise a
battery, or a suitable AC power adapter, or preferably a combination of the
2 0 two where ordinarily power is provided from an AC main with battery
power as a backup.
In Fig. 1 the radio block 10 is shown to include a main antenna
20 connected to a radio frequency duplexer 30 which is in turn connected
to a transmitter block 40 and a RF switch 35. RF switch 35 is connected to a
receiver block 50 and selects between an input "a" from RF duplexer 30 or
from an input "b" from a bandpass filter 27 which is in turn connected to a
diversity antenna 25. An example of the antenna arrangement is described
in commonly assigned Great Britain Patent Application GB9616174.0
naming Kitchener as inventor, entitled "An Antenna Arrangement" the
3 0 disclosure of which is hereby incorporated by reference. However, the
actual antenna arrangement is not crucial to the invention. In particular,
CA 02219096 1997-10-24
two or more antennas can be used, which can be spatially separated, or as
an alternative, the signals received can have different polarizations.
Furthermore, a combination of space and polarization diversity can be
used.
Both the receiver 50 and transmitter 40 of the radio block 10 are
connected to the RF modulator/demodulator and baseband interface block
70 of the baseband block 60. The baseband block 60 also includes a suitable
baseband Digital Signal Processor (DSP) 80 and a suitable baseband
microcontroller 90 which is in turn connected to a TIA port 100. The TIA
(test interface adapter) port is used to communicate with a data terminal
(e.g., a personal computer) using a TIA unit in order to set the terminal
into various states and carry out commands and/or procedures either for
testing or verification. The Texas Instruments TCM 4300 ARCTIC
(Advanced RF Cellular Telephone Interface Circuit) chip is suitable for
block 70 and the TI TDMA-3 DSP is suitable for the baseband DSP 80. The
baseband microcontroller 90 is a suitable air interface microprocessor
which supports the call processing requirement of the IS-54B with
extended protocol along with associated memory (e.g. RAM, ROM,
EEPROM).
Most of the communications between the radio and baseband
blocks takes place via the baseband interface 70. However, as can be seen, a
control link, labelled ANTSEL allows the baseband microcontroller 90 to
toggle the RF switch 35.
In operation, a communication signal is received at both the
main antenna 20 and the diversity antenna 25, where the signal is suitably
filtered either by the RF duplexer 30 or the band pass filter 27 respectively.
The RF switch 35 determines which of the signals, a or b, are
downconverted by the receiver block 50 based on suitable diversity
selection process, as discussed below.
3 0 The selected signal is downconverted to a suitable IF signal by
the receiver block 50. The receiver block 50 also measures the received
CA 02219096 1997-10-24
signal strength and sends a message to the RF demodulator and baseband
interface block 70 as to the received signal strength indication (RSSI) as is
known in the art. The RSSI is then sent to the baseband microcontroller
90. The baseband DSP 80 also determines the bit error rate (BER) which is
also forwarded to the baseband microcontroller. The baseband
microcontroller 90 is used to process layers 1 to 3 of the communication
protocol stack, to manage the control of the RF radio block 10 and the
baseband block 60, and also carries out user interface functions.
The invention is primarily concerned with how to use the
10 RSSI and BER measurements in order to determine how to toggle RF
switch 35 in order to select which antenna is to be used for reception.
The diversity selection process according to one embodiment of
the invention will be discussed with reference to the flow chart of Fig. 2
and the timing reference shown in Fig. 3. Fig. 2 is a flow chart of the
operations carried out by the base band microcontroller 90, according to
this embodiment of the invention which is configured for an IS-54 B,
TDMA-3 system which supports MAHO. For example, Fig. 2 represents
the steps carried out according to software programs stored in an associated
memory (not shown) of the microcontroller. It should be noted that this
2 o example is described for a TDMA-3 system wherein each frame has 6
timeslots and each frame is divided into two half frames with each half
frame having one receive timeslot of interest per user. In the example
illustrated in Fig. 3, there are 6 timeslots per frame (3 timeslots for each
half frame), wherein timeslots 1 and 4 are of interest to the terminal. Fig. 2
can best be understood with reference to the following definitions:
A = antenna A (main antenna);
B = antenna B (diversity antenna);
SW = RF switch 35 for selecting between A and B;
3 o thrO1 = parameter defining a suitably high RRSI level such that switching
antennas is not required;
CA 02219096 1997-10-24
"BEST" is a variable that defines which antenna is selected as the best
antenna for the next half frame;
RSSI_A = the RSSI value measured on antenna A; and
RSSI_B = the RSSI value measured on antenna B.
In Fig. 2, step 250 represents an initialization step wherein the
RF switch 35 is set to receive from the main antenna (A) 20 via RF
duplexer 30. Furthermore, this system sets as a default the main antenna
A as the best antenna for the next half frame. The system will then receive
10 data, which can be either control information or voice or data from a
traffic channel. After reception during the appropriate time slot, for
example, time slot 1, is completed there is a period of time before the
system transmits data. In this embodiment, which is suitable for IS 54B
which is a TDMA/FDD system, this period of time is approximately 3.7
milliseconds (labelled as idle A at 320 in Fig. 3). During this period, as
shown at step 252 in Fig. 2, the synthesiser of the receiver 50 is tuned to an
another frequency in order to make RRSI (receive signal strength
indication) measurements for mobile assisted hand off (MAHO) according
to IS-54 B requirements. This allows the MAHO measurements to be made
20 during the idle time when there is no reception or transmission. This has
the advantage of preventing transmit leakage through the duplexer during
the MAHO measurement, thus allowing an absolute measurement to be
made which is unaffected by the transceiver's own transmissions. The
system then retunes the synthesiser to the receive channel frequency and
the RF switch is toggled to receive from the main antenna A if it is not
already as shown at step 253.
The next step as shown at 256 involves a measurement of the
RSSI as received at antenna A. Advantageously, this step allows for the
measurement of the RSSI of the received channel while the unit is also
3 o transmitting data on the transmit channel as can be seen at 345 in Figure 3. The diversity measurements can be made during transmission because
CA 02219096 1997-10-24
any leakage typically affects both RSSI measurements equally. Thus, the
diversity selection process is able to use relative measurements of the two
antenna RSSI's, rather than the absolute measurements generally required
for MAHO. However, as shown in Figure 3, measurements are not made
during the transmit slot turn-on period 343 due to the transient affect on
one of the RSSI measurements.
In order to compare the RSSI between antennas the RF switch
is toggled to select the signal originating from the diversity antenna as
shown at step 260. The RSSI value from this signal is then measured at
10 263. If necessary, the RF switch is reset back to the value associated with
the BEST variable at step 266 for reception. In other words, the RSSI
measurements are made on both antennas and then the last selected BEST
antenna is again selected for the next receive time slot where data is then
demodulated at step 268.
After the RSSI measurements are made from both antennas,
both values are compared against a threshold value as shown at step 270.
If both RSSI_A and RSSI_B are above the threshold value thrOl then both
antennas are receiving a sufficiently strong signal. In this case, the value
of BEST is unchanged and the system continues to receive with the
20 current antenna. However, if either RSSI_A or RSSI_B are below the
threshold value thrOl then the system compares the signal strengths
received from each antenna as shown at step 280. If RSSI_A is greater
than or equal to RSSI_B then, as shown at step 285, the BEST antenna is
set to A. Then, at step 287 the RSSI_A value is sent to the DSP to adjust
internal DSP automatic gain control (AGC) in order to avoid bit hits which
result when the maximum input level of the baseband demodulator is
exceeded. The cycle then continues in the next frame. However, if the
RSSI_B was greater than RSSI_A, then correspondingly, as shown at steps
290 and 295, the BEST antenna is set for the diversity antenna B and the
3 o RSSI received at antenna B is sent to the DSP, prior to the cycle continuing in the next frame.
CA 02219096 1997-10-24
As stated above, and as can be seen in Fig. 3, the measurements
of the RSSI at antenna A and at antenna B occur during the transmission
time slot. As described, the selection process determines which antenna
should be used for the next receive slot. As an alternative, given sufficient
processing in the time available, the process can determine the antenna to
be used for reception during the current half frame. For example, referring
to Fig. 3, the terminal receives once every half frame (i.e., every 3 time
slots, e.g., timeslot 1, 4,1, etc.) for a TDMA-3 system. It is preferable,
assuming there is sufficient time during Idle B, to select the BEST antenna
10 prior to the start of timeslot 4. However, unless there are very rapid fades,it has been found sufficient to revert to the current antenna for timeslot 4,
so that any change in the BEST antenna occurs for the next timeslot (i.e.,
timeslot 1 in the next frame).
If, as shown in this implementation, the diversity RSSI
measurements are made during the transmit burst then, duplexer 30 must
have sufficient filtering to prevent the transceiver's own transmissions
from interfering with its diversity antenna RSSI measurements. However,
for systems with sufficient processing capabilities, it is possible to conduct
the RSSI measurements during the idle times between the transmit and
20 receive bursts.
Thus an advantage of the above diversity selection process
involves making the diversity measurements as to the BEST antenna to be
used, in conjunction with the MAHO measurements, between the end of
one receive slot (i.e., a timeslot of interest) and the beginning of the next
receive slot. Thus, all of the data in the receive slot is received using the
same antenna. Thus bit hits due to RSSI measurements or resulting from
switching antennas during a receive slot are avoided.
Preferably, both RSSI_A and RSSI_B represent the average
RSSI power levels during the time periods that the measurements are
3 o made. Furthermore, in order to avoid "ping ponging" between antennas,
a running average of the RSSI power level over more than one half frame
CA 02219096 1997-10-24
can be used. Furthermore the switching rate can be constrained by a
hysteresis value M, preventing the switching between antennas unless the
outcome of the RSSI comparison is the same for M sequential half frames,
where M can take on the values of 1, 2,... 255. These refinements are
advantageous in situations where the fading is slow or where a
demodulator is sensitive to frequent changes to antennas. Such situations
typically occur if the two antennas are spaced sufficiently far apart that they
see different signal path delays, or in situations where the two antenna
signal levels are often drastically different. In these circurnstances, either
10 hysterisis, averaging or both can be applied.
The above described diversity selection process is particularly
suited to switch antennas in order to avoid multi path fading. According
to another embodiment of the invention, the terminal can switch
antennas, even if the currently selected antenna has a higher average
RSSI, in order to improve poor reception caused by co-channel
interference (CCI).
Figure 4 is a flow chart of the steps carried out by the baseband
microcontroller, according to software programs stored in an associated
memory (not shown) of the microcontroller, according to an embodiment
20 of the invention which uses both RSSI and BER to determine the antenna
selection. In this embodiment, the diversity selection process selects
which antenna is used based on testing whether co-channel interference is
present, as well as testing for signal level. In the embodiment described,
the bit error ratio (BER) is used as a part of the test for CCI performance.
As an alternative to measuring the BER, or as a supplement, the system
can measure the Coded Digital Verification Colour Code (CDVCC) parity
check and confirm that it has been correctly decoded.
Fig. 4 can best be understood with reference to the following
definitions:
A = antenna A (main antenna);
CA 02219096 1997-10-24
B = antenna B (diversity antenna);
SW = RF switch 35 for selecting between A and B;
thrO1 = parameter defining a suitably high RRSI level such that switching
antennas is not required;
"BEST" is a variable that defines which antenna is selected as the best
antenna for the next half frame;
RSSI_A = the RSSI value measured on antenna A; and
RSSI_B = the RSSI value measured on antenna B.
"CURRENT" is a variable that defines the current antenna, A or B.
10 "OTHER" is a variable that defines the other antenna not in use, either B
or A.
thr20 = a BER threshold parameter set to identify whether or not the BER
is OK
"M" is a hysterisis parameter (M=7 for BER hysterisis and M=O for RSSI
based switching. M=7 is chosen for BER to ensure interferer persists and
not spurious. O<M<255)
"Hysteresis Cnt" is a variable used as a counter
"CCI Hold Cnt" is a variable used as a counter
"DF" is a flag used to disable one portion of the code (testing purposes)
2 0 "JF" is a flag used to alter operation of algorithm to adjust for different
durations of interfering signals
thrO1 = - 50 dBm (point at which diversity is enabled (range is -120 < thrO1
< - 40)
thr20 = 0.5 % BER (value for deciding if 13ER is OK), range is 0% < thr20 <=
8%)
thr21 = -90 dBm (value above which RSSI is considered OK, ran~e is -120 <
thr21 < -40 )
"Z" is a parameter that defines the length of the CCI hold interval in half-
frame increments
CA 02219096 1997-10-24
In Fig. 4, steps 402, 404, 406, 408 represent initialization steps
wherein the RF switch 35 is set to receive from the main antenna (A) 20
via RF duplexer 30. Furthermore, this system sets as a default the main
antenna A as the best antenna for the next half frame. The system will
then receive data, which can be either control information or voice or data
from a traffic channel. After reception during the appropriate time slot,
for example, time slot 1, is completed there is a period of time before the
system transmits data. In this embodiment, which is suitable for IS-54B
which is a TDMA/FDD system, this period of time is approximately 3.7
10 milliseconds (labelled as idle A at 320 in Fig. 3). MAHO measurements, if
desired, are preferably made during this idle period. The system then
retunes the synthesiser to the receive channel frequency and the RF switch
is toggled to receive from the main antenna A if it is not already as shown
at step 420.
The next step as shown at 423 involves a measurement of the
RSSI as received at antenna A. In order to compare the RSSI between
antennas the RF switch is toggled to select the signal originating from the
diversity antenna as shown at step 426. The RSSI value from this signal is
then measured at 429. If necessary, the RF switch is reset back to the value
2 o associated with the BEST variable at step 432 for data reception. In other
words, the RSSI measurements are made on both antennas and then the
last selected BEST antenna is selected for the receive time slot where data
is then demodulated at step 435. This RSSI information is saved for use at
steps 452 and 465.
The next steps 439 to 449 inclusive determine whether or not
the BER is OK. "BER is OK" is a state which is defined based on parameters
set and the current BER as demodulated in step 435. If BER is OK the
selection process uses the faster RSSI based switching, initiated at step 465
to combat multipath fading. If BER is not OK, then a slower switching
3 o process based on signal quality measurements over a number of frames is
initiated in step 452 to determine if CCI exists.
CA 022l9096 l997-l0-24
14
Steps 439 to 449 test for determining if BER is OK. After the
half frame boundary at step 439, the measured BER from the previous
demodulation, step 435, is compared to a threshold value thr20 at step 440.
If the BER is less than thr20 then "Hysteresis Cnt" is reset to zero in step
443. If it is not less than thr20 then Hysteresis Cnt is incremented by 1. In
step 449 the value of Hysteresis Cnt is then compared to M to decide if BER
is OK. These steps 439 to 449 inclusive implement a hysteresis function
such that the BER needs to be greater than or equal to thr20 for "M"
consecutive half frames before selection process determines that BER is
not OK. This prevents incorrect decisions from being made due to burst
errors by testing only for non-transient poor BER conditions.
If BER is not OK the slower multiframe based switching is
initiated at step 452. Figure 4c illustrates the steps for implementing the
portion of the selection process that makes use of RSSI information and
BER information to decide if CCI exists. If so, then antenna switching is
driven based on BER such that an alternate antenna, which may have a
better SIR due to a different multipath combination of signal and
interferer, can be selected.
At step 480 the current antenna RSSI is compared to thrO1 in
order to determine if RSSI is OK. If the RSSI is OK (i.e., RRSI_Current
greater than thr21) then CCI is assumed to exist since the signal strength is
good but the BER is not OK, recognizing that RSSI measurements measure
signal plus interference power. Step 486 compares the OTHER RSSI to
thrO1 to decide if it is OK. If OTHER RSSI is OK then the algorithm
branches to step 492 and then switches to the OTHER antenna in step 495.
The hysteresis counter CCI Hold in step 492 prevents excessive toggling
between antennas from occurring by maintaining the selected antenna for
Z half frames. This holding on the selected antenna is implemented in
conjunction with steps 455, 458 and 470.
3 o At step 486, if the outcome is that OTHER RSSI is not OK, then
the algorithm assumes that there is no advantage to switch. Therefore the
CA 02219096 1997-10-24
current antenna remains the best antenna (step 498), thus eliminating
unnecessary switching to the other antenna. Excessive or unnecessary
toggling between antennas is known to contribute to performance
degradation in typical "blind switching" or "switch and stay" algorithms.
At step 480, if CURRENT RSSI is not OK, the poor BER,
determined in step 449 is likely caused by low signal strength of the
received signal. The selection process then proceeds to step 483 where the
OTHER RSSI is tested. If the RSSI of the other antenna is also not OK (i.e.,
below thr21) then in step 498, the current antenna remains the best
10 antenna. If the OTHER RSSI is OK then a switch to the other antenna is
initiated in step 495. Optional step 489 provides an option to jump out of
the slow BER based switching (if JF =1) by turning any previously set hold
off by setting CCI Hold Cnt = 0. This flexibility allows tailoring of the
selection process to interference and signal variations found in various
propagation environments.
With further reference to step 449, if BER is OK, then a hold
check is performed at step 455 to determine whether the current antenna
should be maintained due to a recent CCI based switch to it. (This duration
is controlled by parameter Z). Step 458 decrements this counter each half
20 frame for which a "hold" condition still exists.
At step 455, if CCI Hold Cnt = 0, then both BER is OK and there
are no recent switches due to poor BER (i.e., no hold condition exists) so
the half frame by half frame RSSI based switching is enabled at step 465 to
combat multipath fading. Note that optional step 461 allows for a disable
flag DF to be set in order to disable the RSSI based switching, for example,
in order to allow testing of the BER based switching only.
At step 465 the multipath fading based switching process is
initiated and the selection process proceeds as shown in Figure 4D.
RSSI measurements from both antennas are compared against
3 o a threshold value as shown at step 500. If both are RSSI_A and RSSI_B are
above the threshold variable thrO1 then this implies that both antennas
CA 022l9096 l997-l0-24
16
are receiving a sufficiently strong signal, therefore no switching of the
antenna is required. In this case, the value of BEST is unchanged and the
system continues to receive with the current antenna. However, if either
RSSI_A or RSSI_B are below the threshold variable thrO1 then the system
compares the signal strengths received from each antenna as shown at step
505. If RSSI_Current is greater than or equal to RSSI_Other then, as
shown at steps 510 and 515, the BEST antenna is set to the current antenna.
Optionally, the RSSI_Current value is sent to the DSP to adjust internal
DSP automatic gain control (AGC) in order to avoid bit hits which result
10 when the maximum input level of the baseband demodulator is exceeded.
This step is shown to be defaulted off as it is typically only required in
conditions when the two signals from the two antennas are either fading
rapidly or have significant time dispersion between them.
The cycle then continues in the next frame. However, if the
RSSI_Other was greater than RSSI_Current, then correspondingly, as
shown at steps 520 and 525, the BEST antenna is set for the other antenna
and, optionally, the RSSI received at the other antenna is sent to the DSP.
As shown at step 530 the Hysteresis Cnt variable is reset to zero prior to the
cycle continuing in the next frame as a switch to the other antenna has
20 been selected.
One advantage of the combined RSSI and BER based switching
is that it will automatically adjust to different channel conditions, whether
degradations are dominated by multipath fading or by CCI.
Note that the preferred embodiment of the design which uses a
main antenna and diversity antenna based on polarization diversity
effectively provides microscopic diversity. In this case the BER based
algorithm relies on the likelihood that the multipath fading on the
interferer is different on the alternate antenna. (Note the mean signal and
interferer levels will be approximately the same over a duration of a few
3 o minutes in the fixed access case).
CA 02219096 1997-10-24
An alternative embodiment of this invention is to use a
remote main antenna whereby the distance separating the main and
diversity antenna are larger and hence the mean signal and interferer
levels are likely to be different due to shadowing and hence the BER based
algorithm can combat longer term poor SIR at the subscriber terminal.
As described, the selection process determines which antenna
should be used for the next receive timeslot. Depending on the frame
structure and processing time of the system, this can be either the timeslot
of interest in the current (half) frame or the timeslot of interest in a
10 subsequent (half) frame.
Fig. 5 is a flow chart illustrating the conceptual steps according
to one embodiment of the invention, for which Fig. 4 represents a specific
implementation. For example, steps 483 and 486 determine whether or not
RSSI_Other is OK as a rough estimate of whether the other antenna signal
quality is OK. In Fig. 4, this test makes use of the fact RSSI_Other is
known. However, other tests, for example testing BER may be preferable
in some circumstances especially if additional DSP processing is available.
Step 600 represents the beginning of each frame. In this
embodiment, the first step 610 involves evaluating whether a quality
20 indication of the signal received on the current antenna has satisfied a
particular condition. Preferably this is determined by evaluating the bit
error rate (BER), CDVCC, or both, during each frame and determining
whether the quality has satisfied a threshold over M frames. Checking the
quality over M frames has the advantage of avoiding excessive switching
due to a transient drop in quality. If the test at 610is not satisfied, the
system evaluates whether it is beneficial to select the other antenna due to
poor quality on the current antenna. As shown at step 620 an evaluation
as to whether the signal received on the other antenna satisfies a second
condition is made. The determination of 620 can be made by evaluating
3 0 the RSSI of the other antenna against an established threshold. However,
in some circumstances where there is sufficient processing available, a
CA 022l9096 l997-l0-24
18
determination of both the quality and strength of the other antenna can be
made. In any event, if the other antenna fails to satisfy the criteria there is
no advantage in switching and the selection of the current antenna is
maintained as shown at step 660. However, if the evaluation of step 620
indicates that the other antenna does satisfy the condition, it is beneficial
to select the other antenna for reception. Therefore, the other antenna is
selected as the best antenna at step 630. A determination is then made as
to whether the poor quality results from a weak signal, for example due to
a fade condition, or alternatively whether the poor quality results from co-
10 channel interference. This determination is made at step 640 byevaluating whether the RSSI satisfied a third condition, for example
whether the RSSI exceeded a minimum threshold during the current
frame. If the current RSSI failed to satisfy the condition, this is evidence
that the quality was poor due to co-channel interference rather than a poor
signal strength. Therefore, a hold condition is established at step 650
preventing subsequent switching of the selected antenna for a specified
number of frames.
A determination as to whether a hold condition exists is made
to prevent switching back to an antenna that was previously deemed to
20 have poor SIR. As shown Fig. 5, the evaluation as to whether the hold is
set is made at step 670. A person skilled in the art should note that this
hold evaluation could alternatively be made prior to step 610. In any
event, if a hold is set, a hold counter, which counts the number of frames
since the hold condition was set, is decreased 680 until the hold condition
no longer exists 685. In this situation, the selection process continues the
next frame with the current antenna selected as the best antenna.
However, if no hold has been set, then at step 690, a
determination is made as to which antenna is receiving the stronger signal
and that antenna is selected for the next frame. The process then
3 0 continues into the next frame. In the embodiment described, the quality
indication (e.g. BER) is tested over "M" consecutive frames. A test for poor
CA 02219096 1997-10-24
19
quality could alternatively be satisfied if x out of y consecutive timeslots
fail to pass the threshold (e.g., poor quality if 3 out of 4 consecutive frames
are poor).
Note that alternative antenna arrangements can be used. For
example, a pair of optional remote diversity antennas which are selected
by a separate RF switch according to an external communication link from
the controller can be used.
Numerous modifications, variations and adaptations may be
made to the particular embodiments of the invention described above
10 without departing from the scope of the invention, which is defined in the
claims.
