Telephonic communications – Echo cancellation or suppression – Using digital signal processing
Reexamination Certificate
1999-09-07
2002-08-27
Isen, Forester W. (Department: 2644)
Telephonic communications
Echo cancellation or suppression
Using digital signal processing
C379S406010, C379S406060, C379S406090, C379S406030, C379S406040, C379S406140
Reexamination Certificate
active
06442274
ABSTRACT:
BACKGROUND OF THE INVENTION
The present invention relates to an echo canceling method or apparatus that can simultaneously cancel echoes occurring in plural transmission lines.
Regarding the techniques of canceling echoes leaking from the transmitter side to the receiver side on the 4-line side of a 2-4 wire hybrid transformer, an echo canceler disclosed in “Adaptive Signal Processing”, 1985. Practice-Hall Inc., USA (reference 1—see Appendix 1) is known.
An echo canceller uses an adaptive filter with tap coefficients equal to or exceeding the impulse response length of the echo path. The echo canceller generates a pseudo echo (echo replica) corresponding to the transmitted signal and then suppresses the echo leaking from the transmitter side to the receiver side in the 2-4 wire hybrid transformer. Each tap coefficient of the adaptive filter can be adapted by taking the correlation between the transmitted signal and an error obtained by subtracting the echo replica from a mixed signal containing the echo and the received signal.
As a typical coefficient adaptation algorithm for the adaptive filters, the normalized LMS (NLMS) disclosed in “Adaptive Filters”, 1985, Kulwer Academic Publications, USA (reference 2—see Appendix 2) is known.
In actual communication lines, plural subscriber lines are multiplexed to further improve the communication capacity for efficiency. In a such case, echo cancellers that remove echoes in the 2-4 wire hybrid transformer are provided corresponding to the number of lines to be multiplexed.
FIG. 31
shows a configuration of a prior art of multiplexed echo cancellers where the number of multiplex is three. In the first channel, a transmission signal is input to the transmission signal input terminal
1
and then is transmitted to the transmission line via the transmission signal output terminal
2
. The transmission signal is further transmitted to the 2-line side of the 2-4 wire hybrid transformer
3
. However, the mismatch in impedance causes a part of the transmission signal to leak as an echo into the receiving side. The echo is supplied to the subtracter
5
via the received signal input terminal
4
. The adaptive filter
86
receives the input signal
700
input to the input terminal
1
and then performs the convolution of the input signal with a corrected coefficient value of the adaptive filter
86
based on the error signal
702
which is the output of the subtracter
5
, thus, creating an echo replica
701
. The subtracter
5
subtracts the echo replica
701
given by the adaptive filter
86
from the echo leaking into the receiver side and then transmits the subtracted result to the received signal output terminal
6
. The subtracted result is further fed back to the adaptive filter
86
as the error signal
702
for coefficient updating.
FIG. 32
is a block circuit diagram illustrating a configuration of the adaptive filter
86
. The adaptive filter
86
has (N−1) delay elements including the delay elements
20
l
to
20
N−1
each of which delays the transmission signal
700
. The number of taps, including the tap with a delay of zero, is N. The adaptive filter
86
further has N coefficient generators
310
l
to
310
N
to generate tap coefficients thereof. N delayed samples, or outputs of the delay elements, are supplied to the corresponding coefficient generators
310
l
to
310
N
and the multipliers
40
l
to
40
N
.
The multipliers
40
l
to
40
N
respectively multiply tap coefficients output from the coefficient generators
310
l
to
310
N
by delayed samples from corresponding delay elements and then outputs the results to the adder
8
. The adder
8
adds all the results from the multipliers
40
l
to
40
N
and then outputs the sum as the echo replica
701
. The memory
105
supplies a step size used for coefficient adaptation to with the coefficient generators
310
l
to
310
N
.
The coefficient generator
310
i
(i=1, 2, . . . , N) has the configuration shown in FIG.
33
. The multiplier
31
multiplies the error
702
by a step size. The multiplier
32
multiplies the resultant product by a delayed signal from each of the delay elements
20
l
to
20
N−1
. The output of the multiplier
32
represents a coefficient correction amount. The adder
33
adds the output of the multiplier
32
and a coefficient value stored in the memory
34
and then feeds back the resultant sum to the memory
34
. The value delayed by the memory
34
becomes the coefficient value after updating.
The configuration of echo cancellers in the second and third channels shown in
FIG. 31
is similar to that in the first channel. The operation of each element is similar to that in the first channel. Hence, the duplicate description will be omitted here.
The conventional echo canceller for multiplexed lines requires computations which increase in proportion to the number of multiplexed lines. This results from the fact that the characteristics of signals input to the multiplexed line are not considered.
SUMMARY OF THE INVENTION
The present invention is made to solve the above-mentioned problems.
The objective of the invention is to provide an echo canceling method suitable for multiplexed lines, which requires a small amount of computations.
Furthermore, the objective of the present invention is to provide an echo canceling apparatus suitable for multiplexed lines, which requires a small amount of computations.
The echo canceling method or apparatus for multiplexed lines of the present invention is characterized in that the convergence degrees of adaptive filters in plural channels are mutually compared and the coefficient updating of adaptive filters whose convergence is leading is suppressed.
Specifically, the echo canceling apparatus has a controller-that evaluates a set of information regarding time-varying step sizes received from adaptive filters respectively connected to transmission lines,. thus, respectively supplying coefficient update suppression signals to the adaptive filters.
Moreover, the echo canceling apparatus has a controller that receives a set of information regarding tap coefficient positions, in the area of the tapped delay line where tap coefficients are most concentrated, from adaptive filters respectively connected to transmission lines, and evaluates variations of the information, thus, supplying coefficient update suppression signals to the respective adaptive filters.
REFERENCES:
patent: 4507747 (1985-03-01), Houdard et al.
patent: 5323459 (1994-06-01), Hirano
patent: 5371789 (1994-12-01), Hirano
Borrallo et al ; “On the Implementation of a Partitioned Block Frequency Domain Adaptive Filter (PBFDAF) for long Acoustic Echo Cancellation”; Int. J. of Signal Processing, vol. 27, No. 3. Jun. 1992; Pub: Elsevier, New York; pp. 301-315.*
Gitlin et al; Data Communications Principles; 1992, Pub. Plenum Press, New Your, Chapter 9, Section 9.3; pp. 619-623.*
Bernard Widrow, et al. “Adaptive Signal Processing” Prentice Hall, pp. 338-347.
Michael L. Honig, et al. “Adaptive Filters: Structures, Algorithms, and Applications” Kluwer Academic Publishers, pp. 54-57.
Akihiro Hirano, et al. “A Noise-Robust Stochastic Gradient Algorithm With an Adaptive Step-size Suitable For Mobile Hands-free Telephones” 1995, IEEE, pp. 1392-1395.
Yutaka Hiratani, et al. “A Noise-Robust Echo Canceller on V830 Multimedia Risc Processor Integrated into a Car Navigation System” 1998, IEEE, pp. 1753-1756.
V. John Mathews, et al. “Stochastic Gradient Adaptive Filters With Gradient Adaptive Step Sizes” 1990, IEEE, pp. 1385-1388.
Akihiko Sugiyama “An Interference-Robust Stochastic Gradient Algorithm With a Gradient-Adaptive Step-Size” Apr. 1993, IEEE International Conference on Acoustics, Speech and Signal Processing.
Akihiko Sugiyama “An Adaptive-Step LMS Algorithm With Coefficient Value Evaluation (Alcove)” 1991, A-75, C&C Systems Research Laboratories.
Tyseer Aboulnasr, et al. “A Robust Variable Step-Size LMS-Type Algorithm: Analysis and Simulations” IEEE Transactions On Signal Processing, vol. 45, No. 3, Mar. 1997.
T. Creasy, et al. “A Rob
Dickstein Shapiro Morin & Oshinsky LLP.
Isen Forester W.
NEC Corporation
Singh Ramnandan
LandOfFree
Method and apparatus of canceling echoes in multi-channel does not yet have a rating. At this time, there are no reviews or comments for this patent.
If you have personal experience with Method and apparatus of canceling echoes in multi-channel, we encourage you to share that experience with our LandOfFree.com community. Your opinion is very important and Method and apparatus of canceling echoes in multi-channel will most certainly appreciate the feedback.
Profile ID: LFUS-PAI-O-2955904