Title of Invention

JOINT CHANNEL PARAMETER ESTIMATION

Abstract A method and apparatus for jointly estimating channels used to propagate desired and interfering signals received by a wireless receiver. A blind parameter estimator in the wireless receiver includes a forward parameter estimator and a backward parameter estimator that each include an equalizer and a channel estimator for generating forward and backward channel estimates, respectively, for the desired and interfering signal channels. In one embodiment, the blind parameter estimator includes independent forward and backward parameter estimators adapted to generate independent forward and backward channel estimates, respectively, where a final channel selector is adapted to select the final channel estimates based on a comparison between error metrics associated with the forward and backward channel estimates. In another embodiment, the blind parameter estimator includes serial per-survivor processing forward and backward parameter estimators adapted to use the backward channel estimates as the final channel estimates for the desired and interfering signal channels.
Full Text BACKGROUND OF THE INVENTION
The present invention generally relates to processing received wireless
communication signals and particularly relates to using joint channel estimation to cancel
interfering signals from a received wireless communication signal.
In response to an ever-increasing demand for wireless services, wireless providers
continue searching for new ways to increase the capacity of current wireless networks.
Because interference within a system limits capacity, one way to increase system capacity is
to reject or cancel interference using digital signal processing techniques. Interference
cancellation can be performed by jointly demodulating the desired and interfering signals.
One type of joint demodulation requires the received signal to contain synchronized desired
and interfering signals. When the desired and interfering signals are synchronized, the
training symbol period of the interfering signal roughly overlaps the training symbol period of
the desired signal. As a result, the joint demodulation process may exploit the overlapping
training sequences present in the received signal to jointly estimate the desired and
interfering signal channels.
However, the desired and interfering signals received by the wireless communication
device are not always synchronized, e.g., in current Time Division Multiple Access (TDMA)
cellular radiotelephone systems like Global System for Mobile communications (GSM),
Enhanced Data Rates for GSM Evolution (EDGE), and Digital-Advanced Mobile Phone
Service (D-AMPS). As a result, the training sequence of the interfering signals does not
overlap the training sequence of the desired signal, which negatively impacts the
performance of the synchronized joint channel estimation process discussed above.
To address this problem, the wireless industry continues to explore methods of joint
channel estimation that do not require the desired and interfering signals to be synchronized,
and therefore, do not require knowledge of the training symbols associated with the
interfering signal.
SUMMARY OF THE INVENTION
The present invention comprises a method and apparatus that uses a known symbol
sequence associated with a desired signal to jointly estimate parameters of the radio
channel that propagate the desired and interfering signals received by the wireless
communication device. According to the present invention, a blind parameter estimator in a
receiver of the wireless communication device includes a forward parameter estimator and a
backward parameter estimator. Each of the forward and backward parameter estimators
includes an equalizer and a channel parameter estimator that generate forward and
backward parameter estimates, respectively, for each of the desired and interfering signal


channels based on iterative forward/backward recursions through the equalizer. Based on
at least one of the generated forward and backward parameter estimates, the blind
parameter estimator is adapted to generate final parameter estimates for the desired and
interfering signal channels.
According to one exemplary embodiment, the blind parameter estimator comprises
independent forward and backward parameter estimators that independently perform
forward and backward recursions to generate forward and backward parameter estimates,
respectively, for the desired and interfering signal channels. In some embodiments, one or
both of the forward and backward parameter estimators may comprise a per-survivor
processing (PSP) parameter estimator that is adapted to generate the forward/backward
parameter estimates from tentative parameter estimates produced at each stage of the
forward/backward recursionfor a plurality of hypothesized survivor paths.
In either case, each of the forward and backward parameter estimators
independently perform one or more iterations of each of the forward and backward
recursions, where the forward/backward parameter estimates resulting from a current
iteration operate as initial forward/backward parameter estimates for a subsequent iteration.
After the final iteration, a final parameter selector is adapted to select the final parameter
estimates for each of the desired and interfering signal channels based on a comparison
between forward and backward error metrics associated with the final iteration's forward and
backward parameter estimates, respectively.
According to another exemplary embodiment of the present invention, the blind
parameter estimator comprises serial first and second PSP parameter estimators. In this
embodiment, the first parameter estimator is adapted to generate tentative first parameter
estimates in a first direction at each stage of the first recursion based on the received signal
and initial first parameter estimates for a plurality of hypothesized forward survivor paths.
Based on the tentative first parameter estimates, the first parameter estimator is adapted to
generate a set of first parameter estimates. Similarly, the second parameter estimator is
adapted to generate tentative second parameter estimates in a second, opposite direction at
each stage of the second recursion based on the received signal and initial second
parameter estimates for a plurality of hypothesized second survivor paths, where the
recently generated set of first parameter estimates operate as the initial second parameter
estimates. Based on the tentative second parameter estimates, the second parameter
estimator is adapted to generate a set of second parameter estimates.
In this embodiment, the exemplary blind parameter estimator performs one or more
ations, where a single iteration includes the first recursion in the first direction followed by
recursion in the second direction. In some embodiments, a single iteration may


include additional recursions in alternating directions. When subsequent iterations are
performed, the set of second parameter estimates generated by the second parameter
estimator in a current iteration operate as the initial first parameter estimates for the first
parameter estimator in a subsequent iteration. After the final iteration, the blind parameter
estimator uses the set of second parameter estimates generated in the final iteration as the
final parameter estimates for the desired and interfering signal channels.
BRIEF DESCRIPTION OF THE DRAWINGS
Figure 1 illustrates a block diagram of one exemplary wireless communication device
according to the present invention.
Figure 2 illustrates a block diagram of one exemplary wireless receiver in the
wireless communication device of Figure 1.
Figure 3 illustrates a block diagram of one exemplary unidirectional parameter
estimator.
Figure 4 illustrates a block diagram of one exemplary PSP unidirectional parameter
estimator.
Figure 5 illustrates a block diagram of a multi-directional parameter estimator
according to the present invention.
Figure 6 illustrates a block diagram of one exemplary multi-directional parameter
estimator according to the present invention.
Figure 7 illustrates a block diagram of one exemplary PSP multi-directional
parameter estimator according to the present invention.
Figure 8 illustrates another block diagram of one exemplary PSP multi-directional
parameter estimator according to the present invention.
Figure 9 illustrates another block diagram of one exemplary PSP multi-directional
parameter estimator according to the present invention.
Figure 10 illustrates a generic block diagram of a parameter estimator.
DETAILED DESCRIPTION OF THE INVENTION
Figure 1 illustrates one exemplary wireless communication device 100, such as a
mobile terminal, base station, or any other wireless device that includes a wireless receiver.
As used herein, the term "mobile terminal" may include cellular telephones, satellite
telephones, personal communication services (PCS) devices, personal data assistants
(PDAs), palm-top computers, laptop computers, pagers, and the like. Further, those skilled
in the art should note the present invention is described in one or more exemplary
embodiments relating to GSM/EDGE wireless communication networks, but such


descriptions are not limiting. Therefore, it should be understood that the present invention
has a broad range of applicability, including other wireless communication standards
including, but not limited to, Universal Mobile Telecommunication System (UMTS), TIA/EIA-
136, Code Division Multiple Access (CDMA), cdmaOne, cdma2000, and Wideband CDMA.
As illustrated, wireless communication device 100 includes an antenna 102, an
optional antenna switch 104, a transmitter 110, a receiver 120, a system controller 150, and
a user interface 160. In one exemplary embodiment, system controller 150 stores and
executes program instructions that control transmitter 110, receiver 120, and optional
antenna switch 104. Further, system controller 150 interfaces the communication
electronics (transmitter 110 and receiver 120) with the user interface 160. When optional
antenna switch 104 connects transmitter 110 to antenna 102, transmitter 110 transmits
wireless communication signals according to a predefined communication standard.
Similarly, when optional antenna switch 104 connects receiver 120 to antenna 102, receiver
120 receives and processes a received wireless communication signal according to a
predefined communication standard. It will be appreciated by those skilled in the art that
when the transmitter 110 and the receiver 120 of wireless communication device 100 are
decoupled in frequency, antenna switch 104 may be omitted.
Figure 2 illustrates one exemplary embodiment of wireless communication receiver
120. As shown in Figure 2, the receiver 120 receives wireless communication signals that
represent a combination of a desired signal and one or more interfering signals, where the
desired signal includes a sequence of training symbols known to the receiver 120. It will be
appreciated by those skilled in the art that the interfering signals may use the same or a
different modulation type as the desired signal. For purposes of the following discussion, the
interfering signal may include any type of interference, such as co-channel interference,
adjacent channel interference, etc., and is asynchronous with the desired signal. Therefore,
the training symbols and the timing associated with the interfering signal are unknown to
receiver 120. However, those skilled in the art will appreciate that the invention described
below may also be applied to a received signal having synchronous desired and interfering
signals.
Receiver 120 includes front-end 122, initialization circuit 124, filtering circuit 126
(optional), demodulator or joint demodulator 128, and blind parameter estimator 130. Front-
end 122 processes a desired signal and one or more interfering signals received at antenna
102 to provide a sampled signal rn to initialization circuit 124 and blind parameter estimator
130 using filters, amplifiers, analog-to-digital converters, mixers, etc., as understood in the
art.


Initialization circuit 124 generates initial channel estimates for each of the desired
signal channel and the interfering signal channel. The initial channel estimates for the
desired signal channel may be generated by any known means. For example, because
receiver 120 has prior knowledge of the training symbols associated with the desired signal,
initialization circuit 124 may use the Least Squares (LS) estimation process to generate
initial channel estimates for the desired signal channel using the known training symbols.
Alternatively, in interference-limited environments, initialization circuit 124 may use the
Constant Envelope (CE) method, which models the interference as a constant envelope
process.
However, because receiver 120 does not have prior knowledge of the training
sequence or the timing of the interfering signal, the above-described methods used to
generate the initial channel estimates for the desired signal channel cannot be used to
generate the initial channel estimates for the interfering signal channel. Instead, initialization
circuit 124 may use a predefined value as the initial interfering channel estimate. For
example, initialization circuit 124 may set all of the channel taps of the initial interfering
channel estimate to zero. Alternatively, initialization circuit 124 may set one of the channel
taps to a small value greater than zero, such as 0.1, and set the remaining channel taps to
zero. Typically, the initialization circuit 124 positions the non-zero channel tap proximate the
middle of the initial channel estimate.
After initialization circuit 124 generates the initial channel estimates for the desired
and interfering signal channels, blind parameter estimator 130 uses joint channel estimation
to generate the final channel estimates for the desired and interfering signal channels, as
discussed further below. While not required, receiver 120 may also include filtering circuits
126, as shown in Figure 2, to further reduce noise affecting the signals to be demodulated by
demodulator 128. For example, filtering circuits 126 may include a whitening filter to whiten
the noise. Filtering circuits 126 may also include a pre-filter to compact the signal energy in
the leading channel taps of the whitened channel estimates. In any event, demodulator 128
demodulates the received signal rn using the desired signal channel estimate generated by
blind parameter estimator 130 and optionally filtered by filtering circuit 126 to estimate the
desired signal. It will be appreciated by those skilled in the art that demodulator 128 may be
a conventional demodulator or a joint demodulator that jointly detects the symbols of the
desired user and interferer(s).
As discussed above, the blind parameter estimator 130 of the present invention
jointly estimates the desired and interfering signal channels to reduce interference, and
therefore, to improve signal quality and network capacity. To better understand the


operation of blind parameter estimator 130, a brief explanation of the operation of
conventional blind parameter estimators is provided herein. A conventional blind parameter
estimator 132, shown in Figure 3, includes an equalizer 134 and a channel estimator 136. In
general, equalizer 134 comprises any known equalizer, such as a Viterbi equalizer, a
decision feedback equalizer, etc., that uses the sampled received signal rn, the known
training symbols of the desired signal, and the initial channel estimates provided by
initialization circuit 124 to detect the symbols of the interfering signal that overlap the training
period of the desired user's training symbols. It will be appreciated that equalizer 134 may
use any known algorithm, such as the Viterbi algorithm, the List Viterbi algorithm, the
Generalized Viterbi algorithm, etc. More particularly, at each stage of a recursion through
the equalizer, equalizer 134 generates an error signal for each surviving path. Equations (1)
through (4) mathematically illustrate this process.
Equation (1) represents the sampled received signal rn:

where wn represents white Gaussian noise, sk,n represents the symbols of the kth signal
(herein, k=1 represents the desired signal and k=2 represents an interfering signal), ck,i
represents the Ith channel tap of the kth signal channel (spanning
symbols), and represents the modulation rotation angle for user k. As used herein k = 1
represents the desired signal and k = 2 represents the interfering signal(s). The received
signal of Equation (1) may also be written in vector notation as:

In operation, equalizer 134 may be any known equalizer, such as a Maximum
Likelihood Sequence Estimation (MLSE) equalizer, a Decision Feedback Sequence
Estimation (DFSE) equalizer, etc. For example, supposing that the desired signal symbols
are known to the receiver, an MLSE equalizer finds the interfering signal symbols that
minimize the error metric:

In Equation (3), represents an error signal given by:


where represent the channel taps for the kth signal at sample time n (for the first time
instant the channel taps are obtained from initialization circuit 124), and are the
hypothesized symbols for the interfering signal produced by the equalizer 134 in, for
example, a trellis.
Using the interferer symbol decisions associated with the surviving path having the
best error metric Λn, channel estimator 136 jointly estimates the desired and interfering
signal channels. For example, channel estimator 136 may comprise a Least Means Squares
(LMS) estimator that estimates the desired and interfering signal channels according to:

where is the error signal associated with the interferer signal decisions , µk is the
LMS step size for user k, and * indicates the Hermetian transpose operation. At the first
stage of the equalizer trellis, channel estimator 136 uses the initial channel estimates
provided by initialization circuit 124. However, after the first stage, channel estimator 136
provides updated channel estimates at each stage. During each subsequent stage,
equalizer 134 uses the updated channel estimates generated by the channel estimator 136
for the current stage to proceed to the next stage. Once the equalizer trellis has been
traversed, the updated channel estimates generated by channel estimator 136 from the final
stage of the equalizer trellis are output as the final channel estimates for the desired and
interfering signal channels. While the above is described in terms of LMS channel
estimation, it will be appreciated by those skilled in the art that other channel estimation
processes may also be used, such as Recursive Least Squares (RLS).
Figure 4 illustrates another conventional parameter estimator 138 that may be used
to jointly estimate desired and interfering channel estimates to reduce interference. The
parameter estimator 138 of Figure 4 is a per-survivor processing (PSP) parameter estimator
that includes equalizer 134 and multiple channel estimators 136. The PSP parameter
estimator 138 includes a channel estimator 136 for each surviving path or state in the
equalizer trellis. It will be appreciated that the parameter estimator 138 of Figure 4 may also
represent a multiple survivor processing (MSP) parameter estimator, where multiple channel
estimators 136 are coupled with each survivor path or state in the equalizer trellis.
In any event, the channel updates for PSP LMS parameter estimation may be given
by:



where are per-survivor channel estimates for the desired and interfering signal
channels obtained by the channel estimator 136 at each stage of the recursion through the
equalizer trellis for each hypothesized surviving path, are hypothesized interfering
signal symbols obtained from equalizer 134 for each hypothesized surviving path, and is
the corresponding error signal. After the final stage, the parameter estimator 138 of Figure 4
selects the channel estimates associated with the best error metric as the final channel
estimate for each of the desired and interfering signal channels.
The parameter estimators 132, 138 shown in Figures 3 and 4, respectively, may be
used for either forward or backward recursions through the equalizer trellis. Further, while
not explicitly shown, either parameter estimator 132, 138 may perform multiple parameter
estimation iterations through the equalizer trellis to generate the final channel estimates for
the desired and interfering signal channels, as well as desired and interfering signal symbol
estimates. In this scenario, the initial channel estimates used by parameter estimator 132,
138 in the first iteration are the initial channel estimates provided by initialization circuit 124.
However, subsequent iterations may use the channel estimates generated by the preceding
iteration as the initial channel estimates for the subsequent iteration. Multiple iterations, and
how they relate to the present invention, are discussed further below.
Turning now to Figures 5-10, various embodiments of the present invention will be
discussed. Figure 5 illustrates a generic embodiment of the inventive blind parameter
estimator 130. According to the present invention, blind parameter estimator 130 includes a
forward parameter estimator 140 and a backward parameter estimator 144. Forward and
backward parameter estimators 140 and 144 may comprise any parameter estimators that
use a forward or backward recursion, respectively, through the equalizer trellis to perform
joint channel estimation, such as the parameter estimators 132 and 138 shown in Figures 3
and 4, respectively, and described above. Generally, the forward and backward parameter
estimators 140 and 144 use known training symbols associated with the desired signal to
jointly estimate the desired and interfering signal channels. However, it will be appreciated
by those skilled in the art that any sequence of symbols associated with the desired signal
and known to the receiver may be used.
Each of the forward and backward parameter estimators 140 and 144 include an
equalizer 134 and at least one channel estimator 136 that generates forward and backward
channel estimates, respectively, for each of the desired and interfering signal channels
based on iterative forward/backward recursions through the equalizer trellis. Based on at

least one of the generated forward and backward channel estimates, blind parameter
estimator 130 generates final channel estimates for the desired and interfering signal
channels. Further, as mentioned above, blind parameter estimator 130 may also provide
symbol estimates for the interfering signal and the desired signal.
Figure 6 shows one example of a blind parameter estimator 130 according to the
present invention. Blind parameter estimator 130 includes forward parameter estimator 140,
optional forward centering circuit 142, backward parameter estimator 144, optional backward
centering circuit 146, and final channel estimate selector 148. In this embodiment, the
forward parameter estimator 140 and the backward parameter estimator 144 operate
independently to generate forward and backward channel estimates, respectively, based on
the known training symbols and initial forward and backward channel estimates. As shown
in Figure 6, forward parameter estimator 140 may iteratively perform one or more forward
recursions through the equalizer trellis to generate forward channel estimates for each of the
desired and interfering signal channels. The initial forward channel estimates used in the
first iteration are the initial forward channel estimates provided by initialization circuit 124.
However, when subsequent iterations are performed, forward parameter estimator 140 uses
the forward channel estimates generated in the current iteration as the initial forward channel
estimates for the subsequent iteration.
Similarly, backward parameter estimator 144 iteratively performs backward
recursions through the equalizer trellis to generate backward channel estimates for each of
the desired and interfering signal channels. Further, the initial backward channel estimates
used in the first iteration are the initial backward channel estimates provided by initialization
circuit 124. However, when subsequent iterations are performed, backward parameter
estimator 144 uses the backward channel estimates generated in the current iteration as the
initial backward channel estimates for the subsequent iteration.
After the final iteration, the forward parameter estimator 140 and the backward
parameter estimator 144 provide the forward and backward channel estimates from the final
iteration, respectively, along with the corresponding forward and backward error metrics to
the final channel estimate selector 148. Final channel estimate selector 148 compares the
forward error metric to the backward error metric to select one of the forward or backward
channel estimates as the final channel estimates. In one exemplary embodiment, final
channel estimate selector 148 selects the channel estimates associated with the best
(lowest) error metric as the final channel estimates.
For the embodiments that include forward centering circuit 142 and/or backward
centering circuit 146, the forward and backward channel estimates generated by the forward
and backward parameter estimators 140 and 144, respectively, are centered according to


conventional means to center the channel response taps of the forward channel estimates
and/or the backward channel estimates. Centering refers to the shifting of the channel
response taps so that the largest tap lies in the middle of the vector of channel response
taps. For example, in the case of a 3-tap channel, the channel response taps may be shifted
to the right/left by one position if the left most/right most tap is the largest tap; the incoming
tap is typically set to zero. It will be appreciated by those skilled in the art that whenever the
desired signal channel estimate is shifted left or right, the received signal also has to be
shifted an equivalent number of positions in the opposite direction. As a result, centering
may also improve synchronization.
In any event, forward centering circuit 142 centers the forward channel estimates
generated by each iteration performed by forward parameter estimator 140 except for the
last iteration. Similarly, backward centering circuit 146 centers the backward channel
estimates generated in each iteration performed by backward parameter estimator 144
except the final iteration. The centered forward/backward channel estimates generated in
each iteration are then used as the initial forward/backward channel estimates for the next
iteration. After the final iteration, the forward centering circuit 142 and the backward
centering circuit 146 are bypassed so that the final channel estimate selector 148 receives
the channel estimates and the corresponding error metric generated by the forward
parameter estimator 140 and the backward parameter estimator 144.
The embodiment shown in Figure 6 may use the parameter estimator 132 shown in
Figure 3 for each of the forward parameter estimator 140 and the backward parameter
estimator 144. However, as discussed above, the blind parameter estimator 130 of the
present invention is not limited to this type of parameter estimator. For example, in one
alternative embodiment of the present invention shown in Figure 7, the forward parameter
estimator 140 and/or the backward parameter estimator 144 may use the PSP parameter
estimator 138 shown in Figure 4.
Another alternative embodiment of the blind parameter estimator 130 of the present
invention is shown in Figure 8. Like the embodiment shown in Figure 7, the blind parameter
estimator 130 of Figure 8 includes a forward parameter estimator 140 and a backward
parameter estimator 144. However, unlike the embodiment shown in Figure 7, the
embodiment of Figure 8 includes serial forward and backward parameter estimators 140 and
144.
In the embodiment illustrated in Figure 8, the forward parameter estimator 140
comprises a PSP parameter estimator, such as the PSP parameter estimator 138 of Figure
4, that generates tentative forward channel estimates at each stage of the forward recursion
through the equalizer trellis for each hypothesized survivor path based en the known training


symbols and the initial forward channel estimates, as described above. Forward parameter
estimator 140 generates a set of forward channel estimates for each of the desired and
interfering signal channels by selecting the tentative forward channel estimates associated
with the surviving paths at the final stage of the forward recursion. The set of forward
channel estimates is then used by the backward parameter estimator 144 as the initial
backward channel estimates for each of the desired and interfering signal channels.
Backward parameter estimator 144 is also a PSP parameter estimator. Therefore,
based on the initial backward channel estimates provided by the forward parameter
estimator 140, the known training symbols, and the received signal, the backward parameter
estimator 144 generates tentative backward channel estimates at each stage of the
backward recursion for each of the hypothesized surviving paths. Backward parameter
estimator 144 generates a set of backward channel estimates for each of the desired and
interfering signal channels by selecting the tentative forward channel estimates associated
with the surviving paths at the final stage of the backward recursion.
As with the embodiments discussed above, the blind parameter estimator 130 of
Figure 8 may perform multiple iterations of the forward and backward parameter estimation
process. However, in this embodiment, a single iteration comprises a forward recursion of
the equalizer trellis in the forward parameter estimator 140 followed by a backward recursion
of the equalizer trellis in the backward parameter estimator 144, where the set of forward
channel estimates generated by the forward parameter estimator 140 operates as the initial
backward channel estimates for the backward parameter estimator 144.
In the first iteration, the forward parameter estimator 140 uses the initial forward
channel estimates provided by initialization circuit 124. However, for subsequent iterations,
the forward parameter estimator 140 uses the set of backward channel estimates generated
by the backward parameter estimator 144 in the current iteration as the initial forward
channel estimates. After the final iteration, final channel estimate selector 148 selects the
backward channel estimate from the set of backward channel estimates having the best
error metric for each of the desired and interfering signal channels.
While Figure 8 illustrates that each iteration of the parameter estimation process
begins with the forward parameter estimator 140 and ends with the backward parameter
estimator 144, it will be appreciated that the process may begin and/or end in any direction.
For example, the serial blind parameter estimator 130 of Figure 8 may use a backward
forward parameter estimation system with a backward parameter estimator 144 followed by
a forward parameter estimator 140. Alternatively, blind parameter estimator 130 may
include multiple parameter estimators, where consecutive parameter estimators perform
opposite recursions through the equalizer trellis. For example, blind parameter estimator


130 may comprise a forward-backward-forward parameter estimator or a backward-forward-
backward parameter estimator. As such, the present invention is not limited to the specific
embodiment shown.
Figure 9 illustrates another embodiment of a blind parameter estimator 130 according
to the present invention. Like the embodiment of Figure 8, the blind parameter estimator
130 of Figure 9 includes serial forward and backward parameter estimators 140 and 144.
However, in the embodiment of Figure 9, the set of forward channel estimates generated by
forward parameter estimator 140 comprises the tentative forward channel estimate for each
of the desired and interfering signal channels associated with the surviving path having the
best error metric. As a result, the set of forward channel estimates used by the backward
parameter estimator 144 as the initial backward channel estimates comprises a single
forward channel estimate for each of the desired and interfering signal channels.
Similarly, the set of backward channel estimates generated by backward parameter
estimator 144 comprises the tentative backward channel estimate for each of the desired
and interfering signal channels associated with the surviving path having the best error
metric. As a result, the set of backward channel estimates used by the forward parameter
estimator 144 in a subsequent iteration as the initial forward channel estimates comprises a
single backward channel estimate for each of the desired and interfering signal channels.
After the final iteration, blind parameter estimator 130 uses the backward channel estimates
generated in the final iteration as the final channel estimates.
When included in blind parameter estimator 130, forward centering circuit 142 and
backward centering circuit 146 function as described above to center the forward and
backward channel estimates for the desired and interfering signal channels provided by
forward and backward parameter estimators 140 and 144, respectively. In this embodiment,
the backward parameter estimator 144 uses the centered forward channel estimates form
the current iteration as the initial backward channel estimates. Further, in all iterations
except the first iteration, forward parameter estimator 140 uses the centered backward
channel estimates as the initial forward channel estimates for the subsequent iteration. After
the final iteration, backward parameter estimator 144 bypasses the backward centering
circuit 146 and uses the backward channel estimates generated in the final iteration by the
backward parameter estimator 144 as the final channel estimates.
While Figure 9 shows a blind parameter estimator 130 including a forward parameter
estimator 140 followed by a backward parameter estimator 144, it will be appreciated that
the blind parameter estimation process may begin and/or end in any direction. Therefore,
like the embodiment of Figure 8, the present invention is not limited to the specific
embodiment shown.


The above describes exemplary blind parameter estimation methods and apparatus
for jointly estimating the desired and interfering signal channels used to propagate the
desired signal channel and the interfering signal channel, respectively. The above-described
methods and apparatus may be used in any wireless communication device, including those
with single or multiple antenna receivers.
The blind parameter estimation process discussed above assumes that there are no
known symbols for any of the received interfering signals. However, the present invention is
not so limited. For example, in situations where the training symbols of some of the
interfering signals are known, the known training symbols may be used to estimate the
corresponding interfering signals according to any conventional means, while the interfering
signals having no known symbols are estimated according to the blind parameter estimation
process described above.
Further, while the above assumes that the receiver 120 has knowledge of the training
symbols associated with the desired signal, such knowledge is not required. When there are
no known training symbols available to the receiver, initialization circuit sets the desired
initial channel estimates and the initial interfering channel estimates to a predefined value as
discussed above. For example, initialization circuit 124 may set all of the channel taps of the
initial and/or desired interfering channel estimate to zero. Alternatively, initialization circuit
124 may set one of the channel taps to a small value greater than zero, such as 0.1, and set
the remaining channel taps to zero. In any event, equalizer 134 uses the sampled received
signal rn and the initial channel estimates provided by initialization circuit 124 to hypothesize
the symbols of both the desired signal and the interfering signal using joint demodulation.
It will also be appreciated that the channel estimates for both the desired and
interfering signal channels may be updated at any desired rate. For example, the channel
estimates may be updated for every received symbol. Alternatively, the channel estimates
may be updated every X symbols. For example, the channel estimates may be updated
ever X = 4 symbols.
In addition, while the above is described in terms of joint channel response
estimation, the present invention may be used to estimate any number of joint channel
parameters. As shown in Figure 10, the channel estimator(s) 136 used in the forward and
backward parameter estimators 140 and 144 may be replaced by channel parameter
estimator(s) 137 that estimate any channel parameter according to the process described
above. These channel parameters may include channel response estimates, frequency
offsets, DC offsets, etc.


The present invention may, of course, be carried out in other ways than those
specifically set forth herein without departing from essential characteristics of the invention.
The present embodiments are to be considered in all respects as illustrative and not
restrictive, and all changes coming within the meaning and equivalency range of the
appended claims are intended to be embraced therein.


WE CLAIM
1. A method of joint channel parameter estimation based on a received signal containing
known symbols associated with a desired signal comprising:
performing a forward recursion through an equalizer trellis to generate a forward channel
parameter estimate for each of a desired signal channel and an interfering signal
channel,
and to generate a corresponding forward error metric based on the received signal and
initial forward channel parameter estimates;
performing a backward recursion through the equalizer trellis, independent from the
forward recursion, to generate a backward channel parameter estimate for each of the
desired signal channel and the interfering signal channel, and to generate a
corresponding
backward error metric based on the received signal and initial backward channel
parameter estimates; and
selecting a final channel parameter estimate for each of the desired and interfering
signal
channels based on a comparison between the forward and backward error metrics.
2. The method of claim 1 further comprising performing the forward and backward
recursions iteratively.
3. The method of claim 2 further comprising:
centering the backward channel parameter estimates; and
using the centered backward channel parameter estimates generated in a current
iteration
as the initial backward channel parameter estimates, respectively, for a subsequent
iteration.
4. The method of claim 2 further comprising using the forward and backward channel
parameter estimates generated in a current iteration as the initial forward and backward channel
parameter estimates, respectively, for a subsequent iteration.
5. The method of claim 2 further comprising:
centering the forward channel parameter estimates; and
using the centered forward channel parameter estimates generated in a current iteration
as
the initial forward channel parameter estimates, respectively, for a subsequent iteration.

6. The method of claim 5 further comprising:
centering the backward channel parameter estimates; and
using the centered backward channel parameter estimates generated in a current
iteration
as the initial backward channel parameter estimates, respectively, for a subsequent
iteration.
7. The method of claim 1 further comprising:
at each stage of the forward recursion, generating a set of tentative forward channel
parameter estimates for each of the desired signal channel and the interfering signal
channel, and generating corresponding tentative forward error metrics based on the
received signal and the forward channel parameter estimates obtained in a previous
stage
of the forward recursion; and
generating the forward channel parameter estimate for each of the desired and
interfering
signal channels and generating the forward error metric based on the set of tentative
forward channel parameter estimates and the corresponding tentative forward error
metrics.
8. The method of claim 1 further comprising:
at each stage of the backward recursion, generating a set of tentative backward channel
parameter estimates for each of the desired signal channel and the interfering signal
channel, and generating corresponding tentative backward error metrics based on the
received signal and the backward channel parameter estimates obtained in a previous
stage of the backward recursion; and
generating the backward channel parameter estimate for each of the desired and
interfering signal channels, and generating the backward error metric based on the set of
tentative backward channel parameter estimates and the corresponding tentative
backward error metrics.
9. The method of claim 1 wherein the known symbols of the received signal comprise
known training symbols associated with the desired signal.

10. The method of claim 1 further comprising hypothesizing interfering signal symbols based
on at least one of the forward and backward channel parameter estimates.
11. The method of claim 1 wherein the forward and backward channel parameter estimates
generated for each of the desired and interfering signal channels comprises forward and
backward channel estimates, and wherein the final channel parameter estimates comprise final
channel estimates for each of the desired and interfering signal channels.
12. The method of claim 1 wherein performing the forward recursion, performing the
backward recursion, and selecting the final channel parameter estimate occurs for every symbol
of the received signal.
13. The method of claim 1 wherein performing the forward recursion, performing the
backward recursion, and selecting the final channel parameter estimate occurs for every X
symbols of the received signal.
14. A blind channel parameter estimator in a wireless communication device to perform joint
channel parameter estimation based on a received signal containing known symbols associated
with a desired signal, the blind channel parameter estimator (130) comprising:
a forward channel parameter estimator (140) adapted to
perform a forward recursion through an equalizer trellis to generate a forward channel
parameter estimate for each of a desired signal channel and an interfering signal
channel, and to generate a corresponding forward error metric based on the received
signal and initial forward channel parameter estimates;
a backward channel parameter estimator (144)adapted to
perform a backward recursion through the equalizer trellis independent from the forward
recursion; and
a backward parameter estimator adapted to generate a backward channel parameter
estimate for each of a desired signal channel and an interfering signal channel, and to
generate a corresponding backward error metric based on the received signal and initial
backward channel parameter estimates; and
a final channel parameter selector (148) adapted to select a final channel parameter
estimate for each of the desired and interfering signal channels based on a comparison
between the forward and backward error metrics.

15. The blind channel parameter estimator (130) of claim 14 wherein the blind channel
parameter estimator (130) is adapted to perform the forward and backward recursions
iteratively.
16. The blind channel parameter estimator (130) of claim 15 wherein in each iteration except
the first iteration, the forward channel parameter estimator is adapted to use the forward
channel parameter estimates generated in a current iteration as the initial forward channel
parameter estimates for a subsequent iteration.
17. The channel blind parameter estimator of claim 16 further comprising a first centering
circuit (142) adapted to center the forward channel parameter estimates after the forward
recursion in each iteration except the last iteration.
18. The blind channel parameter estimator of claim 17 further comprising a second centering
circuit (146) is adapted to center the backward channel parameter estimates after the backward
recursion in each iteration except the last iteration.
19. The blind channel parameter estimator of claim 15 wherein in each iteration except the
first iteration, the backward channel parameter estimator (144)is adapted to use the backward
channel parameter estimates generated in a current iteration as the initial backward channel
parameter estimates for a subsequent iteration.
20. The blind channel parameter estimator of claim 14 wherein the forward channel
parameter estimator(140) comprises a per-survivor forward channel parameter estimator
adapted to generate a set of tentative forward channel parameter estimates at each stage of the
forward recursion for each of the desired signal channel and the interfering signal channel, and
to generate corresponding tentative forward error metrics based on the received signal and the
forward channel parameter estimates obtained in a previous stage of the forward recursion.
21. The blind channel parameter estimator of claim 20 wherein the per-survivor forward
channel parameter estimator is adapted to generate the forward channel parameter estimate for
each of the desired and interfering signal channels, and generates the forward error metric
based on the set of tentative forward channel parameter estimates and the corresponding
tentative forward error metrics.
22. The blind channel parameter estimator of claim 14 wherein the backward channel
parameter estimator (144) comprises a per-survivor backward parameter estimator adapted to
generate a set of tentative backward channel parameter estimates at each stage of the
backward recursion for each of the desired signal channel and the interfering signal channel,

and to generate corresponding tentative backward error metrics based on the received signal
and the backward channel parameter estimates obtained in a previous stage of the backward
recursion.
23. The blind channel parameter estimator of claim 22 wherein the per-survivor backward
channel parameter estimator is adapted to generate the backward channel parameter estimate
for each of the desired and interfering signal channels, and to generate the backward error
metric based on the set of tentative backward channel parameter estimates and the
corresponding tentative backward error metrics.
24. A method of joint channel parameter estimation based on a received signal containing
known symbols associated with a desired signal, the method comprising:
performing a first recursion through an equalizer trellis in a first direction;
at each stage of the first recursion, generating tentative first channel parameter
estimates
for a desired signal channel and an interfering signal channel based on the received
signal
and initial first channel parameter estimates for a plurality of hypothesized forward
survivor paths;
generating a set of first channel parameter estimates for each of the desired and
interfering signal channels based on the tentative first channel parameter estimates;
using the set of first channel parameter estimates generated for each of the desired and
interfering signal channels as initial second channel parameter estimates;
performing a second recursion through the equalizer trellis in a second direction
opposite
from the first direction;
at each stage of the second recursion, generating tentative second channel parameter
estimates for the desired signal channel and the interfering signal channel based on the
received signal and the initial second channel parameter estimates for a plurality of
hypothesized second survivor paths;
generating a set of second channel parameter estimates for each of the desired and
interfering signal channels based on the tentative second channel parameter
estimates;and
using the set of second channel parameter estimates generated for each of the desired
and
interfering signal channels as final channel parameter estimates for the desired and
interfering signal channels.

25. The method of claim 24 further comprising performing said first and second recursions
iteratively.
26. The method of claim 25 further comprising using the set of second channel parameter
estimates generated for each of the desired and interfering signal channels from a current
iteration as the initial first channel parameter estimates for a subsequent iteration.
27. The method of claim 26 wherein generating the set of second channel parameter
estimates for each of the desired and interfering signal channels comprises selecting one
tentative second channel parameter estimate as the set of second channel parameter estimates
generated for each of the desired signal channel and the interfering signal channel based on a
predetermined criteria.
28. The method of claim 27 further comprising centering each of the second channel
parameter estimates selected for each of the desired and interfering signal channels, wherein
using the set of second channel parameter estimates from the current iteration as the initial first
channel parameter estimates for the subsequent iteration comprises using the centered second
channel parameter estimates as the initial first channel parameter estimates for the subsequent
iteration.
29 The method of claim 26 wherein generating the set of second channel parameter
estimates for each of the desired and interfering signal channels comprises selecting two or
more tentative second channel parameter estimates as the set of second channel parameter
estimates generated for each of the desired signal channel and the interfering signal channel
based on a predetermined criteria.
30. The method of claim 24wherein generating the set of first channel parameter estimates
for each of the desired and interfering signal channels comprises selecting one tentative first
channel parameter estimate as the set of second channel parameter estimates generated for
each of the desired signal channel and the interfering signal channel based on a predetermined
criteria.
31. The method of claim 30 further comprising centering each of the first channel parameter
estimates selected for each of the desired and interfering signal channels, wherein using the set
of first channel parameter estimates as initial second channel parameter estimates comprises
using the centered first channel parameter estimates as the initial second channel parameter
estimates.

32. The method of claim 24 wherein generating the set of first channel parameter estimates
for each of the desired and interfering signal channels comprises selecting two or more tentative
first channel parameter estimates as the set of second channel parameter estimates generated
for each of the desired signal channel and the interfering signal channel based on a
predetermined criteria.
33. The method of claim 24 wherein using the set of the second channel parameter
estimates generated for each of the desired and interfering signal channels as the final channel
parameter estimates for the desired and interfering signal channels comprises:
selecting one tentative second channel parameter estimate as the set of second channel
parameter estimates generated for each of the desired signal channel and the interfering
signal channel based on a predetermined criteria; and
using the selected second channel parameter estimates as the final channel parameter
estimates for the desired signal channel and the interfering signal channel.
34. The method of claim 24 wherein the known symbols of the received signal comprise
known training symbols associated with the desired signal.
35. The method of claim 24 further comprising hypothesizing interfering signal symbols
based on at least one of the first and second channel parameter estimates.
36. The method of claim 24 wherein the first and second channel parameter estimates
generated for each of the desired and interfering signal channels comprises first and second
channel estimates, and wherein the final channel parameter estimates comprise final channel
estimates for each of the desired and interfering signal channels.
37. The method of claim 24 wherein the parameter estimation process is executed for every
symbol of the received signal.
38. The method of claim 24 wherein the parameter estimation process is executed every X
symbols of the received signal.
39. The method of claim 24 wherein performing the first recursion comprises performing a
forward recursion and wherein performing the second recursion comprises performing a
backward recursion.
40. A blind channel parameter estimator (130) in a wireless communication device (100) to
perform joint channel parameter estimation based on a received signal containing known
symbols associated with a desired signal, the blind channel parameter estimator (130)
comprising:

a first channel parameter estimator adapted to perform a first recursion through the first
equalizer trellis in a first direction;
to generate tentative first channel parameter estimates at each stage of the first
recursion
for a desired signal channel and an interfering signal channel based on the received
signal
and initial first channel parameter estimates for a plurality of hypothesized first survivor
paths; and
to generate a set of first channel parameter estimates for each of the desired and
interfering signal channels based on the tentative first channel parameter estimates and
wherein the set of first channel parameter estimates operate as initial second channel
parameter estimates;
a second channel parameter estimator
adapted to perform a second recursion through the second equalizer trellis in a second
direction opposite from the first direction
to generate tentative second channel parameter estimates at each stage of the second
recursion for a desired signal channel and an interfering signal channel based on the
received signal and the initial second channel parameter estimates for a plurality of
hypothesized first survivor paths; and to generate a set of second channel parameter
estimates for each of the desired and interfering signal channels based on the tentative
second channel parameter estimates;
wherein the blind channel parameter estimator is adapted to use the set of second annel
parameter estimates generated for each of the desired and interfering signal channelsas
final channel parameter estimates for the desired and interfering signal channels.
41. The blind channel parameter estimator of claim 40 wherein the blind channel parameter
estimator is adapted to perform the first and second recursions iteratively, and wherein the first
channel parameter estimator is adapted to use the set of second channel parameter estimates
generated for each of the desired and interfering signal channels in a current iteration as the
initial first channel parameter estimates for a subsequent iteration.
42. The blind channel parameter estimator of claim 41 wherein the second channel
parameter estimator is adapted to select one of the tentative second channel parameter
estimates as the set of second channel parameter estimates generated for each of the desired
signal channel and the interfering signal channel based on a predetermined criteria.
43. The blind channel parameter estimator of claim 42 further comprising a first centering
circuit adapted to center the second channel parameter estimates selected for each of the

desired and interfering signal channels, wherein the first channel parameter estimator is
adapted to use the centered backward channel parameter estimates as the initial first channel
parameter estimates for the subsequent iteration.
44. The blind channel parameter estimator of claim 41 wherein the second parameter
estimator is adapted to select two or more tentative second channel parameter estimates as the
set of the second channel parameter estimates generated for each of the desired signal channel
and the interfering signal channel.
45. The blind channel parameter estimator of claim 40 wherein the first channel parameter
estimator is adapted to select one of the tentative first channel parameter estimates as the set
of the first channel parameter estimates generated for each of the desired signal channel and
the interfering signal channel based on a predetermined criteria.
46. The blind channel parameter estimator of claim 45 further comprising a second centering
circuit adapted to center the first channel parameter estimates selected for each of the desired
and interfering signal channels, wherein the second channel parameter estimator is adapted to
use the centered first channel parameter estimates as the initial second channel parameter
estimates.
47. The blind channel parameter estimator of claim 40 wherein the first channel parameter
estimator is adapted to select two or more of the tentative first channel parameter estimates as
the set of the first channel parameter estimates generated for each of the desired signal channel
and the interfering signal channel based on a predetermined criteria.
48. The blind channel parameter estimator (130) of claim 40 wherein the second channel
parameter estimator is adapted to select one of the tentative second channel parameter
estimates as the set of the second channel parameter estimates used by the blind channel
parameter estimator (130) as the final channel parameter estimates for each of the desired
signal channel and the interfering signal channel based on a predetermined criteria.
49. The blind channel parameter estimator of claim 40 wherein the first channel parameter
estimator comprises a forward channel parameter estimator (140) and wherein the second
channel parameter estimator comprises a backward channel parameter estimator(144).
50. A wireless communication device (100) comprising:
a transmitter (110) adapted to transmit wireless signals to a different wireless
communication device; and
a receiver (120) adapted to receive a wireless signal containing known symbols
associated with a desired signal, the receiver comprising a blind channel parameter
estimator (130) according to any of claims 41 - 50 or 14-23.

Documents:

01426-kolnp-2007-abstract.pdf

01426-kolnp-2007-assignment.pdf

01426-kolnp-2007-claims1.0.pdf

01426-kolnp-2007-claims1.1.pdf

01426-kolnp-2007-correspondence others 1.1.pdf

01426-kolnp-2007-correspondence others.pdf

01426-kolnp-2007-description complete.pdf

01426-kolnp-2007-drawings.pdf

01426-kolnp-2007-form 1.pdf

01426-kolnp-2007-form 2.pdf

01426-kolnp-2007-form 3.pdf

01426-kolnp-2007-form 5.pdf

01426-kolnp-2007-gpa.pdf

01426-kolnp-2007-international publication.pdf

01426-kolnp-2007-international search report.pdf

01426-kolnp-2007-pct others.pdf

01426-kolnp-2007-pct request.pdf

01426-kolnp-2007-priority document.pdf

1426-KOLNP-2007-ABSTRACT 1.1.pdf

1426-KOLNP-2007-AMANDED CLAIMS.pdf

1426-KOLNP-2007-CORRESPONDENCE 1.3.pdf

1426-KOLNP-2007-CORRESPONDENCE OTHERS 1.2.pdf

1426-KOLNP-2007-CORRESPONDENCE-1.4.pdf

1426-kolnp-2007-correspondence-1.5.pdf

1426-KOLNP-2007-CORRESPONDENCE.pdf

1426-KOLNP-2007-DESCRIPTION (COMPLETE) 1.1.pdf

1426-KOLNP-2007-DRAWINGS 1.1.pdf

1426-kolnp-2007-examination report.pdf

1426-KOLNP-2007-FORM 1 1.1.pdf

1426-kolnp-2007-form 18.pdf

1426-KOLNP-2007-FORM 2 1.1.pdf

1426-KOLNP-2007-FORM 3 1.1.pdf

1426-kolnp-2007-form 3.pdf

1426-kolnp-2007-form 5.pdf

1426-kolnp-2007-gpa.pdf

1426-kolnp-2007-granted-abstract.pdf

1426-kolnp-2007-granted-claims.pdf

1426-kolnp-2007-granted-description (complete).pdf

1426-kolnp-2007-granted-drawings.pdf

1426-kolnp-2007-granted-form 1.pdf

1426-kolnp-2007-granted-form 2.pdf

1426-kolnp-2007-granted-specification.pdf

1426-KOLNP-2007-OTHERS 1.1.pdf

1426-KOLNP-2007-OTHERS.pdf

1426-KOLNP-2007-REPLY TO EXAMINATION REPORT 1.1.pdf

1426-kolnp-2007-reply to examination report-1.2.pdf

1426-KOLNP-2007-REPLY TO EXAMINATION REPORT.pdf

abstract-01426-kolnp-2007.jpg


Patent Number 247776
Indian Patent Application Number 1426/KOLNP/2007
PG Journal Number 20/2011
Publication Date 20-May-2011
Grant Date 18-May-2011
Date of Filing 23-Apr-2007
Name of Patentee TELEFONAKTIEBOLAGET LM ERICSSON (PUBL)
Applicant Address SE-164 83 STOCKHOLM
Inventors:
# Inventor's Name Inventor's Address
1 HAFEEZ, ABDULRAUF 307 PARK YORK LANE, CARY, NC 27519
PCT International Classification Number H04L 25/03
PCT International Application Number PCT/EP2005/010244
PCT International Filing date 2005-09-22
PCT Conventions:
# PCT Application Number Date of Convention Priority Country
1 10/950,868 2004-09-27 U.S.A.