Showing papers on "Kernel adaptive filter published in 1973"

Nth-root Stack Nonlinear Multichannel Filter

Abstract: A nonlinear multichannel filter is developed which appears to be particularly useful for enhancement of seismic refraction and teleseismic array data The basic filter involves the extraction of the Nth root of each element in the matrix forming the data set, where N is any positive integer, and the Nth power of the summation over the channels The filter is effective in reducing random noise, whereas identical signals which are in-phase on all channels are retained at the expense of some distortion The output from this nonlinear filter has far greater resolution in specifying phase velocity than any multichannel linear filter we have employed Examples of theoretical and actual field seismograms are presented after various forms of filtering to illustrate their effectiveness

Unsupervised Tracking of Maneuvering Vehicles

Abstract: The design of adaptive filters for the tracking of high-performance maneuvering targets is a fundamental problem in real-time surveillance systems. As is well known, a filter which provides heavy smoothing can not accurately track an evasive maneuver, and conversely. Consequently, one is led to the consideration of adaptive methods of filter design. This paper presents an improved self-adaptive filter algorithm for on-line solution of the above problem. Basically, this algorithm utilizes the orthogonality property of the residual time series to force the filter to automatically track the optimal gain levels in a changing environment.

On nonlinear filters involving transformation of the time variable

Abstract: A new type of nonlinear filter, called the E -filter, is introduced that involves a transformation of the independent variable of the input function. It is shown how an E -filter can be designed to filter out superimposed "noise" on a signal, leaving the large peaks of the signal unattenuated. Unlike a Iow-pass linear filter, the low-pass E -filter is almost frequency independent and so does not affect the amplitudes of large sharp peaks of the signal. It is shown that the E -filter can be realized in real time and that a wide class of E -filters have a filtering action which is independent of the dc level of the input signal.

The utilization of data measurement residuals for adaptive kalman filtering

Abstract: In recent years the Kalman filter has been utilized extensively for passive target motion analysis (TMA). Unfortunately, in these applications divergence is a common problem. Available methods for eliminating divergence ultimately involve increasing filter sensitivity by discounting the influence of past data. However, this procedure makes the filter more susceptible to random errors; therefore to avoid unnecessarily sacrificing noise performance, adaptive control is required. In this paper the Kalman filter equations are derived and the associated data measurement residuals are examined to determine their suitability for providing adaptive control. An important relationship between the system performance index and the data residuals is established. By exploiting this relationship, pertinent statistical properties of the performance index are deduced and are subsequently utilized as a basis for formulating practical adaptive control criteria. A simulated example is presented to demonstrate divergence (e. g., tracking of a maneuvering target) and significant improvement in performance is noted when adaptive control is appended.

Optimum Signal and Filter Design in Underwater Acoustic Echo Ranging Systems

Abstract: Optimization of the filter, the signal, and the signal and filter jointly are studied in the sonar environment under noise and reverberation limited conditions. The maximization of the receiver output signal-to-interference ratio is used as a performance criterion with unit energy constraint on both signal and filter. In the filter design problem, the optimum filter function is the solution of a linear integral equation. The kernel of the integral equation is a function of the target and medium scattering functions and the reverberation distribution. In the signal design problem, a similar type of integral equation is obtained as in the filter optimization problem. In the joint signal and filter design problem, it is shown that the optimum signal and filter functions are the solutions to a pair of linear integral equations with the largest (SIR)O. Several examples are investigated for different mediums and reverberation distributions with the finite matrix approximation method. An interative technique is used to compute the joint optimization of signal and filter.

Joint Optimization of Receive Filter and Equalizer

Abstract: It is shown that a suitable adaptive receiver for digital communication over noisy dispersive unknown channels consists of an adaptive prefilter followed by an adaptive equalizer. The adaptive receiver is optimized by a minimization of the overall receiver mean-square error. The resultant adaptive receiver requires a dummy equalizer to supply a control signal for adjusting the tap gains of the prefilter.

Polarization Separation by an Adaptive Filter

Abstract: The technique of adaptive filtering as applied to the design of a frequency-reuse receiving antenna which can provide optimum isolation from signals of unwanted polarization.

On sequential refinement schemes for recursive digital filter design

Abstract: A linear synthesis technique is developed for a recursive digital filter which approximates the desired discrete impulse response. Sequential schemes which can be used to refine the preliminary filter design are developed for three distinct cases in which the output sequence number, the filter order, and the number of the numerator coefficients of the transfer function are allowed to vary separately. By use of matrix inversion lemmas, the new sets of filter coefficients are obtained sequentially based on the old estimates and the new data without having to repeat the entire design procedure.

A general analysis for the time-invariant multirate digital filter

Abstract: An analysis is presented for obtaining a general transfer function for a digital filter, with a number of shift sequences in each pulse-repetition interval, as initially described by Fjallbrant. This type of filter has been given the term "multirate filter." The analysis results in a complex convolution integral which is evaluated to produce the transfer function resulting from the multirate processing. Some examples of the use of this technique in certain quantization problems are given.

On the simulation of continuous non-linear filters A comparative study of second- and third-order approximations

Abstract: Among several second-order approximations to the filter of a non-linear system, it is found that the extended Kalman filter is generally the most versatile. The second-order likelihood filter, also known ns the Detchmendy—Sridhar filter is inferior to the ahove and at the same time involves more computation. In the special ease when analytical expressions For the gaussian expectation integrals of the non-linearities can be found, the extended Kalman filter can be further improved by using stochastic linear approximations as suggested by Sunahara. The third-order likelihood filter derived in this paper is more accurate than the above, but calls for considerable storage space and computing time.

Filter design subject to nonlinear side constraints

Abstract: Preliminary results show sequential unconstrained minimization techniques (SUMT) to be a viable filter design method. A nonlinear optimal filter design problem subject to nonlinear side constraints is formulated and solved using SUMT.