Home Home
Azzedine Zerguine, KFUPM (Saudi Arabia)
Maamar Bettayeb, KFUPM (Saudi Arabia)
Colin F.N. Cowan, Queen's University of Belfast (U.K.)
The coefficients of an echo canceller with a near-end section and a far-end section are usually updated with the same updating scheme, such as the LMS algorithm. In this paper we propose a novel scheme for echo cancellation that is based on the minimization of two different cost functions, i.e., one for the near-end section and a different one for the far-end section. Two approaches are addressed and only one of them lead to a substantial improvement in performance over the LMS algorithm when it is applied to both sections of the echo canceller. The proposed scheme is also shown to be robust to noise variations, which is not the case for the LMS algorithm.
ic972313.pdf ic972313.pdf
TOPTomas Gänsler, Lund University (Sweden)
A recursive transfer function estimation algorithm is presented and analyzed. Applications can be found in either electric or acoustic echo cancellation. The proposed algorithm is robust against burst disturbances that are caused by detection misses of double-talk present at the output of the echo path. A frequency-domain technique is used and a robustness function is derived from a criterion that is valid in the application. Analysis of the robust algorithm shows that good performance is to be expected. The performance of the algorithm when operated on real-life speech data in a full duplex communication system is shown by examples. Double-talk detection misses are shown to be well handled by the robust algorithm; yet, convergence rate and variance efficiency are as high as that of a non-robust least squares algorithm.
ic972317.pdf ic972317.pdf TOP
Ling Qin, CENAM/Electronique (France)
Maurice B. Bellanger, CENAM/Electronique (France)
In multicarrier data transmission using filter banks, adaptive equalizers can be introduced in the receiver in every sub-channel, to achieve high bit rates. Following conventional data transmission techniques, two approaches can be envisaged, namely the double sampling equalizer (DSE) and the critical sampling equalizer (CSE). Both schemes are discussed and assessed in the present paper, in the mutlicarrier context. Estimations are given for the length of the equalizers as a function of the channel distortion and of the roll-off factor of the prototype filter in the receiver filter bank. Simulation results associated with two channel models are given to support the theoretical analysis.
ic972321.pdf ic972321.pdf TOP
Kensaku Fujii, Fujitsu Lab. (Japan)
Juro Ohga, Fujitsu Lab. (Japan)
A new type of adaptive algorithm is derived from a first order infinite impulse response (IIR) filter expression of the normalized least mean square (NLMS) algorithm. This new algorithm provides a convergence property similar to that of the recursive least square (RLS) algorithm. Its update equation, however, is extremely simple compared to that of the RLS algorithm. The new algorithm, named sub-RLS algorithm in this paper, can be also derived from the least square (LS) algorithm on an approximation. The prifix sub designates the approximation applied to the LS algorithm for its recursive adaptation. This paper also presents a variation of reducing its processing cost.
ic972325.pdf TOP
Noriyuki Hirayama, Kyoto University (Japan)
Hideaki Sakai, Kyoto University (Japan)
In this paper, an analysis of a delayless subband adaptive digital filter (ADF) structure is presented. In this structure adaptive weights in each subband are computed by the LMS algorithm and then transformed into those in fullband by the Hadamard transform. The conventional subband ADF has transmission delay and aliasing effects associated with the filter bank. However in this manner such defects are avoided while retaining the computational and convergence speed advantages of subband decomposition. In addition the overall transfer function of the novel type of subband ADF is strictly equivalent to that of the Wiener filter for the fullband ADF. Also, a characteristic equation is derived to discuss stability of the adaptation algorithm. Some numerical results show good performance of this scheme.
ic972329.pdf ic972329.pdf TOPShin'ichi Koike, NEC Corporation (Japan)
This paper proposes a new method of theoretical calculation of the expected convergence process for adaptive filters using the Sign Algorithm with digital data as the input reference signal. In the analysis use is made of Gaussian approximated conditional pdf of the error signal to derive a set of difference equations. The results of experiment show sufficient accuracy of the proposed method for practical use, while significantly reducing the computing time in comparison with the previous methods.
ic972333.pdf TOP
Maria D. Miranda, EPUSP (Brazil)
Leonardo Aguayo, EPUSP (Brazil)
Max Gerken, EPUSP (Brazil)
The performance of two minimal QR-LSL algorithms in a low precision environment is investigated. For both algorithms backward consistency and backward stability become guaranteed under simple numerical conventions. They present stable behavior even when excited with ill conditioned signals such as predictable signals. Since the problem of ensuring numerical stability is solved for these algorithms, an investigation about their accuracy is in place. By simulating a channel equalizer configuration it is shown that, for small mantissa wordlengths and forgetting factors not too close to 1, the a priori algorithm performs better due to its dispensing with passive rotations. For forgetting factors very close to one and small wordlengths, both algorithms are sensitive to the accuracy of some well-identified computations. They are compared to an LSL algorithm, based on a priori prediction errors, whose good performance in limited precision environments is known.
ic972337.pdf ic972337.pdf TOP
Kamal Premaratne, University of Miami (U.S.A.)
Mohamed Mansour, ETH, Zürich (Switzerland)
Suppose rate of change of coefficients of a linear time-variant system modeled via a difference equation is restricted. The work presented herein is an attempt at developing an algorithm that determines regions in {coefficient-space} where such a system is guaranteed to be globally asymptotically stable. Such information can be extremely useful in many applications. Some previously published related results are consolidated as well.
ic972341.pdf ic972341.pdf TOP
Ronald DeGroat, UTD (U.S.A.)
Eric Dowling, UTD (U.S.A.)
Darel Linebarger, UTD (U.S.A.)
Dinko Begusic, University of Split (Croatia)
In this paper, we combine spherical subspace (SS) and eigen based updating methods with the affine projection (AP) method to produce a new family of fast SS-AP algorithms that offers additional tradeoffs between computation and adaptive filtering performance. Moreover, the implementation of SS-AP is less complicated than the fast RLS based AP algorithms. For certain applications, e.g., echo cancellation and equalization in digital subscriber loop (DSL) transceivers, SS-AP offers performance that is comparable to AP, but at computational costs that are less than the fast AP algorithms.
ic972345.pdf ic972345.pdf TOP
Sau-Gee Chen, Nat. Chiao Tung University (Taiwan)
Yung-An Kao, Nat. Chiao Tung University (Taiwan)
Ching-Yeu Chen, Nat. Chiao Tung University (Taiwan)
The recently proposed Chen's LMS algorithm costs only half multiplications that of the conventional direct-form LMS algorithm (DLMS). Despite of the merit, the algorithm lacked rigorous theoretical analysis. This work intends to characterize its properties and conditions for mean and mean-square convergences. Closed-form MSE are derived, which is slightly larger than that of DLMS algorithm. It is shown, under the condition that the LMS step size (mu) is very small and an extra compensation step size (alpha) is properly chosen, Chen's algorithm has comparable performance to that of the DLMS algorithm. For the algorithm to converge, a tighter bound for (alpha) than before is also derived. The derived properties and conditions are verified by simulations.
ic972349.pdf ic972349.pdf TOP
Yongbin Wei, Purdue University (U.S.A.)
Saul B. Gelfand, Purdue University (U.S.A.)
James V. Krogmeier, Purdue University (U.S.A.)
In many identification and tracking problems, an accurate estimate of the measurement noise variance is available. A partially adaptive LMS-type algorithm is developed which can exploit this information while maintaining the simplicity and robustness of LMS. This noise constrained LMS (NCLMS) algorithm is a type of variable step-size LMS algorithm, which is derived by adding constraints to the mean-square error optimization. The convergence and steady-state performance are analyzed. Both the theoretical results and simulations show that NCLMS can dramatically outperform LMS, RLS and other variable step-size LMS algorithms in a sufficiently noisy environment.
ic972353.pdf ic972353.pdf TOP
Shivaling Mahant-Shetti, Texas Instruments Inc. (U.S.A.)
Srinath Hosur, Texas Instruments Inc. (U.S.A.)
Alan Gatherer, Texas Instruments Inc. (U.S.A.)
This paper describes a new variant of the least-mean-squares (LMS) algorithm, with low computational complexity, for updating an adaptive filter. The reduction in complexity is obtained by using values of the input data and the output error, quantized to the nearest power of two, to compute the gradient. This eliminates the need for multipliers or shifters in the algorithm's update section. The quantization itself is efficiently realizable in hardware. The filtering section is unchanged. Thus, this algorithm is similar to the sign based variants of the LMS algorithm. However, the complexity of the proposed algorithm is lower than that of the sign-error LMS algorithm, while its performance is superior to this algorithm. In particular, it is close to that of the regular LMS algorithm. The new algorithm also requires much lower area for ASIC implementation.
ic972357.pdf ic972357.pdf TOP
J. Jiang, University of Illinois (U.S.A.)
C.D. Schmitz, University of Illinois (U.S.A.)
B.A. Schnaufer, University of Illinois (U.S.A.)
W.K. Jenkins, University of Illinois (U.S.A.)
Adaptive fault tolerance (AFT) has been developed recently for achieving reliable performance of FIR adaptive filters in the presence of certain types of hardware failures. Previous studies limited the use of AFT to one-dimensional FIR filter structures with simple `stuck-at` fault models that account for errors introduced into the adaptive process by faulty coefficients that cease to adjust properly. This paper considers a broader class of hardware errors which, in addition to stuck-at errors in the multiplier inputs, can be modeled by errors in the outputs of both multipliers and adders. By including a simple circuit that removes erroneous induced constants in the output of the adaptive filter, adaptive fault tolerance can be extended to a broader class of hardware failures.
ic972361.pdf TOP