Telecommunications – Receiver or analog modulated signal frequency converter – With wave collector
Reexamination Certificate
2000-09-20
2004-04-20
Vuong, Quochien B. (Department: 2685)
Telecommunications
Receiver or analog modulated signal frequency converter
With wave collector
C455S295000, C455S296000
Reexamination Certificate
active
06725025
ABSTRACT:
STATEMENT REGARDING FEDERALLY SPONSORED RESEARCH OR DEVELOPMENT
Not applicable.
BACKGROUND OF THE INVENTION
This invention is in the field of wireless communications, and is more specifically directed to methods of cancelling interference among wireless units communicating with the same base station.
The use of wireless communications, for both voice and now also data, has exploded over recent years. Even with the high density of base stations now present in many cities, and with the high capacity provided by code-division multiple access (CDMA) technology, the number of users that may be supported by each base station has become a limiting factor in the system capacity. Indeed, it has been observed that interference from other users within a cell is now the most significant factor limiting cell capacity, even more significant than additive thermal noise.
Interference cancellation (IC) techniques have been developed to improve base station uplink capacity, and thus the number of users that may be served by a base station. In general, IC performed at the base station attempts to remove, from each received user signal, interference from other users before making each data decision (i.e., deciding the value of each symbol or “chip”). Conventional IC techniques can be categorized as either serial or parallel. Serial, or successive, IC processes each user signal in a sequence, cancelling the interference caused by each remaining user in the sequence. Parallel IC simultaneously removes, for each user, the interference from all other users. As may be expected, the delay required for parallel IC is much less than that for serial IC. As such, parallel IC has become quite popular. Additional detail in this regard is provided in Divsalar, et al., “Improved Parallel Interference Cancellation for CDMA”,
IEEE Trans. Communications,
Vol. 46, No. 2 (February 1998), pp. 258-268, and in Moshavi, “Multi-User Detection for DS-CDMA Communications”,
IEEE Communications Magazine
(October. 1996), pp. 124-136. As described in the Divsalar et al. and Moshavi articles, variations of parallel interference cancellation have been proposed.
FIG. 1
illustrates one stage of a conventional parallel interference cancellation (PIC) detector, by way of background. This conventional approach is based on “hard” data decisions. Initial bit estimates d
k
(0) are conventionally derived from a matched filter detector or the like. These initial bit estimates are scaled by amplitude estimator
2
, by the application of an amplitude estimate A for each user (delayed by a delay time T
b
, as are all signals in FIG.
1
). The scaled initial bit estimates are then respread into chip form by spreader
3
, resulting in an estimate of the received signal for each user. These estimates are applied to partial summer
4
which sums, for each user, all of the estimates for all other users (and excluding the user of concern). In other words, for a user i, partial summer
4
produces the partial sum
∑
x
≠
i
s
x
,
where s
x
is the estimate from spreader
3
for a given user x. These partial sums constitute an estimate of the regenerated interference from other users.
The outputs of partial summer
4
are applied to the negative input of a summer
7
associated with each user; the positive input of the ith summer
7
receives the most recently received signal r
i
for its corresponding user i. The output of summer
7
thus provides the an estimate of the received signal r
i
less the regenerated interference from all other users. These estimates after cancellation of the regenerated interference are applied to a matched filter bank
8
, and are then applied to a data decision function
9
for each user, producing the next iterated values of data decisions d
i
(1) for each user (from i=1 to i=k).
These next values of data decisions are then applied to a next stage of the parallel interference cancellation detector, to produce the next succeeding iterated value. The process is then repeated until convergence.
This conventional IC methodology, and its conventional variations, rely on the convergence of the data decisions; however, in practice, no such convergence is guaranteed. In particular, it has been observed, in connection with the present invention, that convergence will occur only if the initial guesses are accurate for a significant fraction of the users being considered. Also, in conventional PIC approaches, convergence tends to reliably occur only if the interference among users is relatively low.
The need for a more robust interference cancellation approach thus exists in the art, particularly as the density of wireless units in the-field continues to increase.
By way of further background, a known sampling technique is referred to in the art as Gibbs sampling. The Gibbs sampler is a Markov-chain Monte Carlo technique in which random variables are drawn from a joint probability density function. Each random variable is supplied with an initial guess, which- need not be particularly accurate. The value of one random variable is then selected from the joint probability density function, using the initial guesses for the other variables. The value of the next variable is then in turn selected from the joint probability density function, given the then-current values of all other variables. Following a transient period that typically involves several passes through the sequence of variables, continued operation of the Gibbs sampler will derive a relatively accurate distribution of values for each variable, from which estimations may be made. Further discussion of the Gibbs sampler may be found in Rajan, et al., “Parameter estimation of time-varying autoregressive models using the Gibbs sampler”,
Electronics Letters,
Vol. 31, No. 13 (Jun. 22, 1995), pp. 1035-1036.
BRIEF SUMMARY OF THE INVENTION
It is therefore an object of the present invention to provide an interference cancellation method and system that is useful for accurately cancelling interference in crowded wireless communications cells.
It is a further object of the present invention to provide such a method and system that may be used in applications in which inter-user interference is significant.
It is a further object of the present invention to provide such a method and system in which convergence does not strongly depend upon the accuracy of the selection of initial conditions.
Other objects and advantages of the present invention will be apparent to those of ordinary skill in the art having reference to the following specification together with its drawings.
The present invention may be implemented into a communications station or device for wireless voice and data communications. In the station, memory is provided for storing samples of multiple interfering signals, in the form of chips or symbols. Random initial guesses are made for the data decisions of each signal. For a selected signal, interference cancellation is then performed by subtracting regenerated interference or by cross-correlation, using the current data decisions (either initial guesses or derived values) for the interfering signals. This provides a “soft” data decision value for the selected signal, which defines a probability distribution from which a random selection process determines a new data decision value. These operations are then repeated, signal by signal, until convergence. Following convergence, the distribution of data decision values for each user generates an estimate of the actual communicated signal. The random selection of a new data decision value from the “soft” data decision, and also the random initial guesses, ensure convergence over a wide range of interference conditions, independently from the values off the initial data decision guesses.
REFERENCES:
patent: 5533012 (1996-07-01), Fukasawa et al.
patent: 5872776 (1999-02-01), Yang
patent: 6161209 (2000-12-01), Moher
Electronics Letters, Jun. 22, 1995, vol. 31, No. 13, Rajan et al., “parameter estimation of time-varying autoregressive mod using Gibbs Sampler”.*
IEEE 0163-6804/96,
Chen Rong
Gatherer Alan
Schmidl Timothy M.
Wang Xiaodong
Brady III Wade James
Chow C.
Neerings Ronald O.
Vuong Quochien B.
LandOfFree
Interference cancellation among wireless units using Gibbs... does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Interference cancellation among wireless units using Gibbs..., we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Interference cancellation among wireless units using Gibbs... will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-3215206