摘要:

The paper sets out to review the area of linear adaptive filters, restricted to the classification of finite impluse response filters. The exact solution to this problem based on the recursive least-squares algorithm is first derived. This algorithm is then degraded to show the evolution of self-orthogonalising adaptive algorithms, and further, to the stochastic gradient search algorithms. Computer simulations are presented to compare and contrast the performance of the algorithms in terms of their convergence bahaviour. The relative complexity and numerical stability of the algorithms is then discussed. Together, these comparisons provide a comprehensive basis on which to base an informed decision on choice of algorithm for any defined application.

DOI:10.1049/ip-f-1.1987.0046

出版商:IEE    

年代:1987

数据来源: IET

摘要:

The tolerances to additive white Gaussian noise of some recently developed near-maximum-likelihood detectors are compared by computer simulation, using models of two telephone circuits and an HF radio link for the transmission path. Each detector is here preceded by an adaptive allpass linear filter that is adjusted to make the sampled impulse response of the channel and filter minimum phase. For every channel, the near-maximum-likelihood detectors achieve a substantial advantage in tolerance to additive white Gaussian noise over a conventional nonlinear equaliser. An approximate estimate of the relative complexities of the different near-maximum-likelihood detectors suggests that the most recently developed of these detectors achieves a much better compromise between performance and complexity than the others, for both types of transmission path. It is also shown that the performances of all detectors tested can be seriously degraded, in operation with an HF radio link, if the adaptive linear prefilter is modified, but still maintained as an allpass network, such that the sampled impulse response of the channel and filter is no longer minimum-phase. Nevertheless, the near-maximum-likelihood detectors maintain their advantage over the nonlinear equaliser, when the same modification is applied to the latter.

DOI:10.1049/ip-f-1.1987.0047

出版商:IEE    

年代:1987

数据来源: IET

摘要:

The paper presents an overview of adaptive filter techniques as applied to the digital coding of speech. Specifically, a model of the human speech production mechanism provides a basis for modelling the speech waveform, and this is exploited to provide efficient means for its digital encoding and transmission. During recent years the expected future demand for high-grade secure mobile communications, and the necessity to conserve bandwidth, have given impetus to an already high level of research activity in this field. The basic model of speech used for analysis has been refined in a number of ways in order to effect a reduction in the required data rate while incurring only a minimal loss of speech quality. The advantages and disadvantages of the various techniques reported in the literature are discussed in the paper. In addition, methods for enhancing voice performance when acoustic noise is present at the microphone are described. This aspect is an important one for a number of applications, particularly mobile communications from moving vehicles. The paper includes sections on digital signal processing hardware developed by Racal to implement the algorithms.

DOI:10.1049/ip-f-1.1987.0048

出版商:IEE    

年代:1987

数据来源: IET

摘要:

The application of Kalman estimation techniques to adaptive filtering problems is briefly reviewed. In particular, three common scenarios are examined in detail. The themes that link these areas are that they involve rigorous or almost rigorous application of the Kalman filter, and they can be modelled by a finite-impulse-response (FIR) filter structure. Two new nonrigorous applications of the Kalman filter to the adaptive infinite-impulse-response (IIR) equaliser problem are presented. The first, a direct approach, involves a simultaneous parameter and state estimation algorithm. In the second, an indirect approach, a Kalman-based IIR system identification algorithm is used to estimate the parameters which define the closed-form optimum IIR equaliser. Finally, the convergence of the two algorithms is compared by computer simulation.

DOI:10.1049/ip-f-1.1987.0050

出版商:IEE    

年代:1987

数据来源: IET

摘要:

Least-mean-square (LMS) algorithms for adaptive infinite impulse response (IIR) filters are proposed which sequentially adjust the parameters of the feedforward and feedback paths so as to minimise the equation error. Both the feed-forward and feedback paths are implemented in cascade second-order sections. However, instead of adapting the coefficients of the second-order sections, the actual root locations are adjusted either in cartesian or in polar co-ordinates. Prequantising the movements of the roots eliminates the need for the arbitrary convergence factors used in normal LMS algorithms. Hardware implementations of the adaptive filters have been investigated, and new filter structures for second-order filters are also proposed. An efficient algorithm for adapting the roots of second-order filters in cascade is given. This algorithm combines a simple gradient approximation together with the prequantisation of root movements. Simulations are conducted to assess the performance of the proposed algorithms.

DOI:10.1049/ip-f-1.1987.0052

出版商:IEE    

年代:1987

数据来源: IET

摘要:

The paper describes new algorithms drawn from the field of control engineering which extend classical Weiner smoothing and prediction concepts to cover a class of two-dimensional (2-D) problems. When combined with a recursive parameter estimator, the 2-D algorithms become self-tuning in nature and provide a powerful new class of adaptive signal processing techniques. Applications include image enhancement and multisensor signal filtering.

DOI:10.1049/ip-f-1.1987.0053

出版商:IEE    

年代:1987

数据来源: IET

摘要:

The suppression of ground clutter returns received by an airborne radar is basically a two-dimensional filtering problem, because the clutter echoes depend on two parameters (velocity, azimuth) instead of velocity only as in case of ground-based radars. This requires two-dimensional sampling (in space and time) of the backscattered echo field, which in practice is fulfilled by a coherent pulse Doppler phased array radar. Previous studies have shown that the space-time clutter covariance matrix of the orderNM×NM(Nis the number of sensors,Mthe number of echoes) has onlyN+Mclutter eigenvalues (instead ofNM), which means that the signal vector space can be reduced. In the paper an adaptive radar receiver with AMTI (airborne MTI) capability is shown and discussed. It is shown that this AMTI receiver approximates well the theoretical optimum; however, the computational expense is drastically reduced. In view of the dramatic progress in the field of microelectronics and algorithms (e.g. systolic arrays), one can expect that such AMTI receivers will be realisable for real-time applications in the near future.

DOI:10.1049/ip-f-1.1987.0054

出版商:IEE    

年代:1987

数据来源: IET

摘要:

Two techniques for the adaptive cancellation of clutter in pulse radars are considered. The first, based on the maximisation of the signal/interference improvement, leads to an eigenvector problem; simpler suboptimal structures according to a straightforward model with a minimum number of parameters are considered. The second, based on the linear prediction of the interference, leads to a whitening (or prediction error) filter. The basic performance of both techniques is evaluated in a typical radar environment of two clutter sources.

DOI:10.1049/ip-f-1.1987.0055

出版商:IEE    

年代:1987

数据来源: IET

摘要:

It is commonly stated that the least-mean-square (LMS) algorithm for adaptive filters is a stochastic version of the steepest descent (SD) optimisation technique, although little work on comparative studies has been reported. The present paper sets out a detailed theoretical and experimental comparison. Equations are derived for the directional variance of the estimated gradient, and these are then experimentally verified by means of a constrained LMS simulation—an ensemble of LMS gradients is computed for a set of points determined by advancing an adaptive system according to the SD gradient. Particular attention is focused on the convergence process, since the LMS algorithm has been criticised for being too slow.

DOI:10.1049/ip-f-1.1987.0056

出版商:IEE    

年代:1987

数据来源: IET

摘要:

A simplification of the well known stochastic iteration algorithm, called the dual sign algorithm (DSA), has been proposed recently for adaptive identification and similar applications. There are two key parameters that must be determined for the efficient use of the algorithm, namely the quantisation levelL2and the thresholdrT. Whereas there exists a simple method for the design of these parameters, in the present paper an optimal design theory is developed by studying the relationship betweenL2and the convergence rate. It is found that there is no simple way to obtain a best value ofrTfor the optimally chosenL2value. Instead, a modification of the DSA is proposed that realises the optimum performance suggested by the theory. Computer simulation is used to support the theoretical results.

DOI:10.1049/ip-f-1.1987.0057

出版商:IEE    

年代:1987

数据来源: IET