LMS Adaptive Filter Research Articles

Adaptive LMS filter with lattice prefilter

A filter structure consisting of a lattice predictor to function as a prefilter and a bank of subfilters using backward predictor errors is presented. It can realise a general transversal transfer function, has a good convergence rate, and is insensitive to the eigenvalue spread, even when a low order lattice predictor is used.

Performance comparison of two implementations of the leaky LMS adaptive filter

The leaky LMS adaptive filter can be implemented either directly or by adding random white noise to the input signal of the LMS adaptive filter. In this correspondence, we analyze and compare the mean-square performances of these two adaptive filter implementations for system identification tasks with zero mean i.i.d. input signals. Our results indicate that the performance of the direct implementation is superior to that of the random noise implementation in all respects. However, for small leakage factors, these performance differences are negligible.

Cellular CDMA overlay situations

An adaptive LMS filter is applied to a cellular CDMA overlay situation in order to reject narrowband interference in a mobile fading channel, ignoring a channel white noise. Both the convergence and steady-state results of the LMS algorithm for the CDMA receiver are presented.

Adaptive fault tolerance for reliable LMS adaptive filtering

Adaptive fault tolerance (AFT) is introduced and applied to finite impulse response (FIR) adaptive filters, Fault-tolerant adaptive filters (FTAFs) that result from using the principles of AFT are analyzed with respect to convergence rate, computational complexity, and hardware overhead. A practical and useful FTAF based on the discrete Fourier transform (DFT) is developed that can tolerate numerous coefficient failures regardless of the input noise statistics. The aims of this paper are to demonstrate the principles of the AFT approach and to motivate future research in this area by drawing attention to similarities between AFT and other methods of fault tolerance that exploit adaptivity, such as neural networks.

Modeling and intellligent chatter control strategies for a lathe machine

In the active chatter control of machine tools, the most effective way to suppress the chatter is to place the actuator as close as possible to the tool tip. However, in practice, it is almost impossible to put the actuator at the same location of the tool tip. Also, in many machines the cutting tools are usually long and may be flexible. Both of these problems pose serious problems in machine chatter control. In order to control the chatter effectively and efficiently, a systematic methodology is proposed in this paper to deal with the modeling and control system design aspects of this challenging problem. Because of the flexibility effect in the tool shaft, conventional active control approaches may not perform in an efficient and effective manner. Here, two advanced control algorithms (LMS adaptive filter and fuzzy CMAC neural network) are proposed to counteract this problem. Experimental results on a lathe machine are also included. Approximately 20 dB reduction in chatter has been achieved.

Fuzzy step-size adjustment for the LMS algorithm

This paper presents a fuzzy step-size LMS adaptive filter. This method allows the adaptive filter to converge faster and to produce smaller steady state errors compared to the fixed step-size LMS adaptive filter. Furthermore, it provides a general framework where the user can incorporate additional safety rules or conditions that will prevent the adaptive algorithm to become unstable and to react to sudden changing conditions more rapidly. Simulation results are presented to compare the performance of the new approach with the fixed step-size LMS algorithm and other commonly used variable step-size LMS algorithms such as those proposed by Harris et al. (1986) and Kwong and Johnston (1992).

Combined self-organising feature map - LMS adaptivefilter for digital co-channel interference suppression

The performance of the conventional LMS adaptive filter used for co-channel interference suppression degrades when the desired signal and the interference have similar characteristics. The authors propose a technique that uses a self-organising feature map to track the interference statistics, which are then used by the LMS adaptive filter to reject the interference. Simulation results and an analysis of the weight error power are presented.

Fully analogue LMS adaptive notch filter in BICMOS technology

A fully analogue adaptive notch filter for meter pulse applications in analogue subscriber line systems is presented. The filter is part of an integrated circuit which includes both low voltage subscriber line interface circuit- (SLIC)- and CODEC-filter functions. The effects of offset, noise, and transient behavior are investigated, and a linear model for stability analysis is derived. Fabricated in a 1 /spl mu/m BiCMOS technology this circuit uses the benefits of accurate SC-filters, low offset comparators, and linear multipliers. Due to the adaptive approach the performance of the filter is independent of process deviations and temperature effects. The filter has been designed for an attenuation of 50 dB @ 12 or 16 kHz, a dynamic range of 103 dB and a settling time of less than 5 ms.

Influence of modeling error on noise reduction performance of active noise control systems using filtered-X LMS algorithm.

This paper investigates the influence of modeling error that is the difference of the characteristic between the secondary path and its model on the behavior of the filtered-X LMS adaptive filter. The equations that describe the behavior of the adaptive filter are presented at first, and the influence of modeling error is considered based on the equations. Both the stability and the noise reduction performance are discussed, and particularly, we discussed in detail the noise reduction performance under modeling error condition that has rarely been considered. In addition, the conditions required for maintaining the noise reduction performance is also presented. The results of the theoretical consideration are confirmed by computer simulation in which the impulse responses measured in a vehicle cabin are used. Through the investigation, it is proved that the algorithm may be stable except for the substantial modeling error case, and the noise reduction performance will be inferior to that under ideal condition in general. These results suggest that we should use the on-line identification system when the secondary path characteristic is time variant.

Optimal on-line modeling of primary and secondary paths in active noise control systems.

The modeling error shown by the difference of the characteristics between the secondary paths and its models will cause the improper adaptation of the filtered-x LMS adaptive filters in active noise control systems. It is necessary to use the on-line modeling systems for avoiding such improper adaptation, and several on-line modeling methods have been proposed hitherto. Especially, one of them is very useful because the method can identify the primary and the secondary path characteristics without using any additional signal. This method, however, will not always provide the proper modeling results, and the conditions for optimizing the results have not been discussed yet. This paper investigates the conditions for optimal on-line modeling of the primary and the secondary path characteristics without using any additional signal for identification. Theoretical analysis produces the specific conditions which the noise control adaptive filter should satisfy. Those results of the theoretical consideration are confirmed by the computer simulation in which the impulse responses measured in a vehicle cabin are used.

Transient and tracking performance analysis of the quantized LMS algorithm for time-varying system identification

This paper investigates the statistical behavior of the finite precision LMS adaptive filter in the identification of an unknown time-varying stochastic system. Nonlinear recursions are derived for the mean and mean-square behavior of the adaptive weights. Transient and tracking algorithm performance curves are generated from the recursions and shown to be in excellent agreement with Monte Carlo simulations. Our results demonstrate that linear models are inappropriate for analyzing the transient and the steady-state algorithm behavior. The performance curves indicate that the transient and tracking capabilities cannot be determined from perturbations about the infinite precision case. It is shown that the transient phase of the algorithm increases as the digital wordlength or the speed of variation of the unknown system decrease. Design examples illustrate how the theory can be used to select the algorithm step size and the number of bits in the quantizer.

Analysis and efficient implementation of partitioned block LMS adaptive filters

The frequency domain block LMS (FBLMS) algorithm used for efficient implementation of adaptive filters whose length may exceed a few thousands of taps is explored further by Asharif et al. (1980). To reduce the latency of such filters, it has been proposed that the adaptive filter convolution sum be partitioned into a few smaller sums and then the FBLMS be applied. We show that the scheme proposed by Asharif et al. suffers from a serious eigenvalue spread problem. We identify the source of this problem and propose a solution to that.

Transition based LMS adaptive filter

The performance of conventional adaptive filters used for cochannel interference suppression degrades when the desired signal and the interference have similar characteristics. The author proposes a technique that makes use of the different data clock phases of the desired signal and the interference to separate the data transitions of the desired signal from those of the interference. This principle is used to modify a least mean square adaptive filter. Simulation results and an expression for the weight error vector are presented.

Equivalent Transfer Function Matrix Realization to the Stability Analysis of a Multiple Channel Filtered-X LMS Adaptive Notch Filter

In this paper, stability studies of a multichannel active noise control system are approximated in terms of equivalent transfer function matrix realization. By considering the presence of the acoustic channel, a multichannel filtered-X LMS adaptive notch filter is used with a specified plant model. It has been shown that the system behaves as a linear time-invariant system with a slow adaption. The convergence coefficient and the estimation error of the plant phase characterize the behavior of the system.

Bit level Systolic Implementation on LMS adaptive filter

Bit level Systolic Implementation on LMS adaptive filter

Multichannel block adaptive filters for real‐time, nonparametric identification of time‐varying plants

Many real‐world applications of active noise control are characterized by transfer functions that vary significantly and unpredictably. The controller’s transfer function models must adapt to these variations. Presented here is a class of adaptive filters that accomplish quasiperiodic system identification updates for feedforward control by using blocks of input–output histories. The algorithms form a one‐dimensional family linking normalized LMS adaptive filters and optimal Wiener filters, and are termed ‘‘block projection’’ algorithms. The system identification proceeds noninvasively, producing nonparametric (FIR) impulse responses. The multichannel generalization and application of these algorithms to system identification, as presented here, is novel. Considerations are described that arise from the algorithms’ implementation in the context of system identification; in particular, the proper weighting of input and output data pairs is discussed. The resulting multichannel control algorithms have been implemented successfully for quieting of a compact distributed source in an anechoic environment, and for local quieting of a diffuse field in a reverberant room. In both cases, error microphones could be moved about, providing a ‘‘mobile quiet zone,’’ and performance was obtained for bandwidths exceeding a decade.

Nonlinear behavior of the adaptive LMS algorithm

This brief demonstrates that the adaptive LMS filter does not necessarily converge in the mean to the Wiener solution through the investigation of the harmonic content of signals in the filter for the case of narrowband signals. A 1 tap filter is examined in detail, and the structure of the estimated and error signals for the filter is investigated for an N tap filter. Simulation results are provided which collaborate and amplify the theoretical findings, showing that the harmonic content of the filter output can be predicted.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Generalized sliding FFT and its application to implementation of block LMS adaptive filters

This paper has two contributions. First, the concept of the generalized sliding fast Fourier transform (GSFFT) as an efficient implementation of the hopping FFT is introduced. Application of the GSFFT is broad and not limited to what has been considered in this paper. The frequency domain block LMS (FBLMS) adaptive filters are then revised, and their implementations for block lengths less than the length of the adaptive filter are studied. The GSFFT and the available pruned FFTs are used to give an efficient implementation of these filters. In the particular case of the block length equal to one, where the FBLMS algorithm reduces to the frequency domain LMS (FLMS) algorithm, it is shown that the latter can be implemented with the order of M complexity, where M is the length of the adaptive filter. >

Bit-serial VLSI implementation of delayed LMS adaptive FIR filters

The delayed least-mean-square (DLMS) algorithm is useful for adaptive finite impulse response (FIR) filtering applications where high throughput rates are required. In the paper, a bit-serial bit-level systolic array based on new schemes for multiplication and inner-product computation is presented to implement DLMS adaptive N-tap FIR filters. The architecture is highly regular, modular, and thus well-suited to VLSI implementation. It has an efficiency of 100% and a throughput rate of one filter output per 2B cycles, where B is the word length of input data. In addition, the proposed array uses a small delay of [(4B+N/2+4)/2B] in the filter coefficient adaptation, where [x] is the smallest integer greater than or equal to x. This ensures that the DLMS algorithm can have good performance under proper selection of the step size. Based on a conservative design technique of static complementary metal oxide semiconductor (CMOS) logic, it is shown that the proposed system can be realized in a single chip for most practical applications.< <ETX xmlns:mml="http://www.w3.org/1998/Math/MathML" xmlns:xlink="http://www.w3.org/1999/xlink">&gt;</ETX>

Local stability of the median LMS filter

Local stability properties of the median LMS adaptive filter are investigated by relating the behavior of the algorithm to the behavior of an associated ordinary differential equation. With independent inputs, the differential equation and the algorithm are shown to be locally stable. On the other hand, several classes of (periodic, nonindependent) inputs are described which cause the differential equation and the algorithm to be unstable about its equilibrium, even in the no disturbance case. This will help delineate those applications for which the median LMS is an appropriate adaptive algorithm. >