Multiplex communications – Communication over free space – Combining or distributing information via time channels
Reexamination Certificate
1998-08-07
2001-06-19
Hsu, Alpus H. (Department: 2661)
Multiplex communications
Communication over free space
Combining or distributing information via time channels
C370S201000, C455S501000
Reexamination Certificate
active
06249518
ABSTRACT:
FIELD OF THE INVENTION
This invention relates to wireless communications, and more specifically, an improved demodulator for cancelling co-channel interference.
BACKGROUND OF THE INVENTION
In cellular communication systems, a demodulator is used at a receiver to extract data symbols such as 1-bits and 0-bits that are modulating a communications signal.
The function of a demodulator is complicated by the addition of additive white Gaussian noise (AWGN) and co-channel interference (CCI) to an information signal as it is transmitted through the flat-fading mobile-radio environment. AWGN is introduced from various background noise sources such as local physical media and electronic movement within communication devices. CCI is introduced when several communication channels in geographically close proximity to one another, using the same or closely spaced frequencies, begin to interfere with each other. It is a goal of the present invention to design a cost-effective and rapid demodulator that extracts modulating symbols from communication signals, notwithstanding the addition of a significant co-channel interference component to the signals as they travel through the mobile radio environment.
There are well known demodulation algorithms, such as the minimum mean square error (MMSE) solution, that allow for the cancellation of significant interference signals from a received signal as long as the number (N) of sources of CCI is less than the number (M) of antennas available to the demodulator. These algorithms are practical for base stations where there are typically a plurality of antennas, but are not for portable handsets and fixed wireless terminals where there is a limitation on physical space and cost. These portable handsets and fixed wireless terminals typically only have one antenna, making CCI cancellation (CCIC) by these known algorithms infeasible.
The nature of the problem can be clarified if the communications signal arriving at the input of a demodulator is analysed. More specifically, if one considers chopping up the signal along its time-axis into a plurality of segments that each correspond to one or a small number of symbols, such a segment, hereinafter the received signal, can be expressed in the following form, hereinafter referred to as equation (1):
r
(
n
)=&agr;
d
s
d
(
n
)+&agr;
I
s
I
(
n
)+
v
(
n
)
The variable “n” is an index used to delineate the different received signals that comprise the communications signal received from the transmitter. “S
d
(n)”, hereinafter the information signal, is the part of the received signal that was modulated by one or more data symbols at the transmitter. Once the demodulator has determined the information signal, it can easily demodulate symbols from it. “s
I
(n)”, hereinafter the co-channel interference signal, is the part of the received signal that was transmitted at the same or closely spaced frequency as the information signal by a transmitter geographically in close proximity to the transmitter of the information signal. This interference signal is the information signal for a communication system that is geographically in close proximity, but is not desired within the received signal for the discussed communication system during demodulation. “&agr;
d
” and “&agr;
I
” are fading coefficients used to model the diminishing or enhancing of the information and interference radio channels respectively caused by changes in physical distances or structures between the transmitter and receiver. These fading coefficients are normally assumed to have a constant value over short periods of time, the duration of the periods being a function of the mobile speed. “v(n)” is the part of the received signal caused by the effects of AWGN, hereinafter the noise signal. The problem can thus be described as isolating the information signal from within the received signal, given that the information signal, interference signal, and the noise signal are unknown.
The solution to the problem is made easier because, assuming a digital modulation scheme is in use, a demodulator always has partial knowledge of the information signal. This knowledge is that the information signal can only be one of x
N
possible signals, where x is the number of symbols modulating each information signal, and N is the number of symbols supported by the modulation scheme in use. For example, if a &pgr;/4-DQPSK (differential quadrature phase shift keying) modulation scheme is in use, the information signal carrying a single symbol would have the following form:
s
d
(
n
)=
s
d
(
n−
1)
e
j&pgr;B/4
,
where B=1,3,5 or 7
In this example, the information signal would have to be one of only four possible signals.
This narrowing of the solution set for the information signal, is important because it allows demodulators to take advantage of cross-correlation detection techniques. These techniques are centred around a method of detecting signals in which the received signal is compared, point to point, with a reference signal that is an estimate of what the received signal should be if modulated by a given symbol. The output of such a detector is a measure of the degree of similarity between the received signal and the reference signal. Demodulators can take advantage of these techniques, by setting the reference signals of a cross-correlation detector to equal each of the x
N
possible information signals that could be within the received signal, and then selecting the reference signal that most closely correlates with the received signal as an estimate of the information signal.
Such cross-correlation detection techniques that ignore the interference and noise signals require that &agr;
d
be greater than &agr;
I
by more than 6 dB and that AWGN not be significant for the results to be within the acceptable reliable range, that being a bit error rate (BER) of less than 2%. If the two fading coefficients are within 6 dB or the AWGN is significant, the differences between a reference signal and a received signal could be just as easily attributable to the effects of CCI, as to differences between the underlying modulating symbols of the signals. Two signals that seem well-correlated may in fact only seem that way due to CCI effects and likewise, two signals that are in fact modulated by the same symbol may be poorly correlated with each other due to CCI effects. The fading coefficients do change over time and just because the average fading coefficients have more than 6 dB of separation does not guarantee that they have such separation for all time periods.
Therefore, in order to function in an environment that is heavily affected by CCI, it is desirable that demodulators that use cross-correlation detection techniques, hereinafter referred to as correlation demodulators, be able to distinguish between differences between received and reference signals that are attributable to CCI, and those that are attributable to their differing underlying modulating symbols. Conventional correlation demodulators, which simply cross-correlate each received signal with all possible reference signals, are unable to make this distinction, and thus select many incorrect reference signals as estimates of information signals when operating in the mobile radio environment.
Some existing correlation demodulators do try to model the effects of AWGN and CCI by making use of history correlation data. History correlation data is a record of the received signals and information signal estimates that have been previously made by the demodulator. An example of such a demodulator is disclosed in U.S. patent application Ser. No. 08/989,265 filed Dec. 11, 1997 by Cui et al and assigned to the assignee of the present application, which describes a correlation demodulator with maximal correlation symbol estimation (MCSE) that uses a weighted average of correlations between previously estimated information signals and their corresponding received signals to select as an estimate of the information signal the reference signal that most closely correlates with the rece
Haszko Dennis R.
Hsu Alpus H.
Nguyen Brian
Nortel Networks Limited
Smart & Biggar
LandOfFree
TDMA single antenna co-channel interference cancellation does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with TDMA single antenna co-channel interference cancellation, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and TDMA single antenna co-channel interference cancellation will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2544040