scispace - formally typeset
Search or ask a question

Showing papers on "Kernel adaptive filter published in 1974"


Journal ArticleDOI
TL;DR: In this article, the authors developed a solution for the resulting complex Wiener filter in terms of complex autocorrelation and cross-correlation functions, and an algorithm for the efficient evaluation of the complex filter weights is also available.
Abstract: The ordinary time or space domain Wiener filter is conventionally obtained for real-valued inputs and real-valued desired outputs. Situations exist for which these variables are complex-valued.It is possible to develop a solution for the resulting complex Wiener filter in terms of complex autocorrelation and crosscorrelation functions. An algorithm for the efficient evaluation of the complex filter weights is also available.

35 citations


Journal ArticleDOI
TL;DR: In this paper, the identification of a process modeled by a stable, linear difference equation of known order is dealt with, where the output is subject to additive observation noise that is identically and independently distributed with zero mean and a constant variance.
Abstract: This paper deals with the identification of a process modeled by a stable, linear difference equation of known order. Its output is subject to additive observation noise that is identically and independently distributed with zero mean and a constant variance. On-line estimators in which the process parameters as well as the process outputs are estimated simultaneously in real time are considered. For improving the stability of such on-line algorithms, a simple adaptive filter for the reference model is proposed. Further, it is shown that inclusion of such a filter relates the resulting bootstrap algorithms to the more general forms of the two stage least squares estimators viz. the k -class, h -class and the double k -class estimators. Effectiveness of the filter in stabilizing the on-line algorithms is demonstrated by using data generated by a fourth-order model.

24 citations


Journal ArticleDOI
TL;DR: In this article, an extension of the Bessel filter is given for which the transfer function is a rational function with finite zeros, and a design example for a second-order all-pass constant time delay filter with linear phase response is given.
Abstract: An extension of the Bessel filter is given for which the transfer function is a rational function with finite zeros A special case is shown to combine the constant magnitude response of the all-pass filter with the linear phase response of the Bessel filter A design example for a second-order all-pass constant time delay filter is given; there is good agreement with theory

23 citations


Journal ArticleDOI
TL;DR: This paper describes an algorithm that is suitable for fast implementations of nonrecursive and recursive digital filters and the memory-speed tradeoff is flexible so that many hardware and software implementations are practical.
Abstract: This paper describes an algorithm that is suitable for fast implementations of nonrecursive and recursive digital filters. High speed is realized at the expense of memory; however, the memory-speed tradeoff is flexible so that many hardware and software implementations are practical. When memory is not limited, the time required to compute a filter output value is independent of the order of the filter.

23 citations


Patent
Yoichi Sato1
18 Mar 1974
TL;DR: In this paper, a self-adaptive equalizer comprises a transversal filter wherein the attenuators are adjusted with reference to the output signals of the filter, and means are provided for producing binary signals representative of the signs of the first filter outputs.
Abstract: A self-adaptive equalizer comprises a transversal filter wherein the attenuators are adjusted with reference to the output signals of the filter. For use in a multilevel data transmission system according to correlative encoding, a first filter removes the correlation encoding from the transversal filter outputs. Means is provided for producing binary signals representative of the signs of the first filter outputs. A second filter encodes the binary signals in accordance with the correlative encoding. The attenuators are adjusted in compliance with the respective products of the difference between one each of the second and transversal filter outputs and the signals derived from those taps of the delay line of the transversal filter to which the relevant attenuators are connected.

14 citations


Proceedings ArticleDOI
01 Nov 1974
TL;DR: In this paper, a generalized quadratic Lyapunov function is constructed to allow several new adaptive algorithms to be synthesized, each improving the degree of stability over previously reported adaptive algorithms.
Abstract: The speeds of the adaptive response of Lyapunov-designed adaptive systems are related to some extent to the degree of stability of the system. A generalized quadratic Lyapunov function is constructed to allow several new adaptive algorithms to be synthesized. The new algorithms each improve the degree of stability over previously reported adaptive algorithms. A root-locus analysis substantiates the improvement in response speed. While the framework for this short paper is that of full-state model-reference adaptation, the synthesis is applicable to systems employing output measurement, such as reduced-order model-reference systems [5] or an adaptive observer [20].

13 citations


Journal ArticleDOI
TL;DR: In this paper, the g-h filter is used as a tracking filter and the target under track is modelled as a constant-velocity system with a correlated random acceleration, equations are derived for the covariances of the filtered and predicted estimates.
Abstract: The g-h filter is often used as a tracking filter. Assuming that the target under track is modelled as a constant-velocity system with a correlated random acceleration, equations are derived for the covariances of the filtered and predicted estimates. These equations are useful to predict the performance of the filter and to select suitable parameters so as to improve performance.

9 citations


Journal ArticleDOI
TL;DR: In this article, a minimum variance linear filter which estimates only a part of the state is presented, and the undesired part is eliminated by a simple linear transformation, and a filter is derived for the new reduced order plant dynamics which contain one-step-correlated noise and a simple delay.

8 citations


Journal ArticleDOI
TL;DR: In this paper, it is shown that the performance of a stacking filter with hyperbola-like moveout curves can be characterized by the analysis of a velocity filter with linear (f-k) transform.
Abstract: The implementation of a stacking filter involves the filtering of each trace with an individual filter and the subsequent summing of all outputs. The actual position of a trace in space as well as certain simultaneous shifts of traces and filter components in time do not influence the process. The resulting output is consequently invariant to various arbitrary coordinate transformations. For a certain useful class of ensembles of non-linear moveout arrival times for signals a particular transformation can be found which transforms a given ensemble into one consisting only of straight lines. It is thus possible to reduce, for instance, the analysis of a stacking filter designed for hyperbola-like moveout curves to the analysis of a velocity filter with linear moveout curves. As the (f—k) transform is a very useful concept to describe a velocity filter, it can consequently be applied to characterize a stacking filter in regard to its performance on input signals with non-linear moveout.

7 citations


Proceedings ArticleDOI
01 Nov 1974
TL;DR: In this paper, the authors compared the performance of a filter that transforms nonlinear measurements in the coordinate system of the state before using in the filter with the conventional extended Kalman (EK) filter.
Abstract: This paper evaluates performance of a filter that transforms nonlinear measurements in the coordinate system of the state before using in the filter. The bias and computational performance of this mechanization are compared with the conventional extended Kalman (EK) filter. The bias in the EK filter is a function of the covariance of the estimation error, while in the state-coordinate-system (SCS) it depends on variance of the transformed measurement noise, which is independent of the estimation error and depends only on the nonlinearity. Thus when the initial covariance of the state is large, the SCS filter has less bias than the EK filter. This reduction in the bias is achieved at a cost of additional computations required to implement the transformation. The improved performance of the filter due to the coordinate transformation is then demonstrated with an example of satellite-aided aircraft navigation.

2 citations



Journal ArticleDOI
TL;DR: This paper looks, in detail, into the use of two variable substitution methods to bound the poles and zeros of the filter for stability and minimum phase considerations.
Abstract: Some recent papers on digital filter design have been concerned with the use of unconstrained minimization techniques, such as the Fletcher-Powell algorithm, to optimize the design to a prescribed frequency domain description. This paper looks, in detail, into the use of two variable substitution methods to bound the poles and zeros of the filter for stability and minimum phase considerations. One of the methods is shown to be superior to the other. Extensions of the work are indicated for two-dimensional recursive filter design and more restrictive bounding on the filter configuration than would normally be considered.

Journal ArticleDOI
TL;DR: In this article, an iteration algorithm is proposed for reducing the nonlinearity bias of an extended Kalman filter, which is a simple modification of the single stage iteration filter as defined by Mehra [3], which uses the same filter gains at each iteration.
Abstract: An iteration algorithm is proposed for reducing the nonlinearity bias of an extended Kalman filter. The algorithm is a simple modification of the single stage iteration filter as defined by Mehra [3], which uses the same filter gains at each iteration. The simplified filter yields bias reduction comparable to the original iteration filter, but requires substantially fewer computations.

ReportDOI
01 Sep 1974
TL;DR: In this article, an adaptive alpha-beta filter used to track targets by search radars is described and the gains alpha and beta are adjusted so as to minimize the mean square error between the target's predicted and measured positions.
Abstract: : An adaptive alpha - beta filter used to track targets by search radars is described The gains alpha and beta are adjusted so as to minimize the mean square error between the target's predicted and measured positions Several examples are given These show that the filter responds rather rapidly to changing environments for a time between samples of 4 s

Journal ArticleDOI
TL;DR: In this paper, the authors used the Kalman filter to detect two of the most dominant errors in an inertial navigation system, the schulering and the earth-rate error components, using the state variables method.
Abstract: Detection of two of the most dominant errors in an inertial navigation system, the schulering and the earth-rate error components, are studied using the Kalman filtering technique. Dynamics of the inertial system are modelled by a combination of two coupled oscillators subjected to additive noise, by using the state variables method, contrary to the technique used in the literature using equations of classical mechanics. In this paper the utilization of the method of incremental coefficients (MIC) algorithm is presented. The variance equation (appearing in the filter) is given by the differential equation in P 12(t, t 0). (The matrices P ij (t, t 0) belong to the MIC algorithm) ; and it was solved by the semi-group property of this algorithm. Both schemes, non-adaptive and adaptive, from an accuracy standpoint are simulated. It is shown that using an adaptive filter approach, the Kalman filter is able to filter out the schulering and the earth-rate oscillations to an acceptable level. The controllability ...

ReportDOI
01 Jul 1974
TL;DR: First, an adaptive filter is shown to be able to correct the gyrocompass output for environmental effects and second, it is demonstrated that a low-pass digital filter is able to reduce the measurement noise in the raw data so that it can be properly passed through the adaptive estimator.
Abstract: : The problem of determining true azimuth from noisy gyrocompass data is presented. First, an adaptive filter is shown to be able to correct the gyrocompass output for environmental effects. Second, it is demonstrated that a low-pass digital filter is able to reduce the measurement noise in the raw data so that it can be properly passed through the adaptive estimator. Computer simulations are presented that indicate that the proposed digital and adaptive filter perform as anticipated.


Journal ArticleDOI
TL;DR: In this paper, the tradeoff between filter sensitivity to the uncertain parameters and minimum mean-square estimation error was evaluated relative to the trade-off between the filter sensitivity and the estimation error.

Proceedings ArticleDOI
01 Nov 1974
TL;DR: The adaptive detector combines the best features of linear matched filtering and hard-limiting receiver structures resulting in a small-signal SNR performance which is an improvement over either of these detectors alone.
Abstract: A detector structure and an adaptive algorithm are proposed for the reception of signals in noise backgrounds possessing broad-tailed probability distributions typical of impulsive noise. The adaptive detector combines the best features of linear matched filtering and hard-limiting receiver structures resulting in a small-signal SNR performance which is an improvement over either of these detectors alone. Furthermore, the adaptive detector is relatively easy to implement and provides efficient performance regardless of the actual noise distribution.