A mean level adaptive detector using nonconcurrent data

Convergence results for a mean level adaptive detector (MLAD) are presented. The MLAD consists of an adaptive matched filter (for spatially correlated inputs) followed by a mean level detector (MLD). The optimal weights of the adaptive matched filter are estimated from one batch of data and applied to a statistically independent batch of nonconcurrent data. The threshold of the MLD is determined from the resultant data. Thereafter a candidate cell is compared against this threshold. Probabilities of false alarm and detection are derived as a function of the threshold factor, the order of the matched filter, the number of independent samples per channel used to calculate the adaptive matched filter weights, the number of samples used to set the MLD threshold, and the output signal-to-noise power ratio of the optimal matched filter. A number of performance curves are shown and discussed     

Estimating matched filter amplitude probability functions byfailure rate analysis

Performance analysis of echolocation systems requires knowledge of the probability density function (pdf) or cumulative distribution function (cdf) of a matched filter output. A method is presented to estimate these and other probability functions from data by estimating the failure rate function, a function employed in reliability theory. The method can also be used to derive approximations to closed-form probability functions. The method is demonstrated using experimental sonar and radar clutter data and a closed-form radar clutter model     

Some Matched Filter Configurations for Infrared Systems

An analysis of the output of three alternative matched filter configurations in an infrared scanning system model is presented. The sensor is corrupted by thermal noise, generation-recombination noise, photon noise, and modulation noise, the latter providing an extreme discoloration in the signal passband. Expressions for the signal voltage density spectrum, signal pulse shape, noise power spectrum, and average noise power at the matched filter output are derived where the integral evaluations attendant to these derivations do not appear elsewhere in the literature. The paper also provides graphical displays of the signal-to-noise power ratio at the filter output versus various system parameters, noise power spectrum out of the matched filter versus ?, and the signal pulse shape out of the filter versus time. Also included are discussions of practically realizable approximations to the matched filters and curve fitting techniques for the signal pulse shape function.     

Ambiguity Functions for Quadratic Phase-Coded Pulse Trains

The matched filter ambiguity function is presented for a burst waveform composed of repeated subbursts, each one of which consists of N pulses in which the phase is varied quadratically from pulse to pulse. The resulting ambiguity function exhibits small residual ambiguities along the delay axis separated by the reciprocal of the pulse repetition frequency (PRF). A cross-ambiguity function is derived which reduces these ambiguities to zero amplitude. A third cross-ambiguity function is presented for a receiver matched to a generalized Hamming weighted repeated quadratic burst. The location in the delay/Doppler plane of the waveform ambiguities for these waveforms is compared with that of an uncoded pulse burst.     

On detection using filter banks and higher order statistics

We suggest a method, based on the use of filter bank and higher order statistics (cumulants), for detection of transient signals. The method first uses a bandpass filter bank, which separates the spectrum of the observed signal into narrow frequency bands. Each subfilter of the filter bank is then followed by a cumulant estimator, and thereby suppressing colored noise. By selecting those subfilters that have large output energies, the filter bank can approximate the behavior of a matched filter. Moreover, no a priori information about the waveform of the signal is needed. The performance of the detector is evaluated by using a simulated signal as well as a measured signal. The presented detector is compared with the optimal matched filter detector.     

A Computationally Efficient Approximation to a Discrete-Time Matched Filter

When the number of filter coefficients is large, the solution of the discrete-time matched filter equation can be computationally difficult. In this paper several techniques are presented for approximating the impulse response of a matched filter without actually solving the matched filter equation. The performance of these approximating filters is analyzed and compared with the performance of the matched filter. It is also shown that an approximation which is best in a mean-squared-error sense is not necessarily best in terms of output signal-to-noise ratio.     

The Susceptivity of MTI Systems to White Noise

The performance of various coherent MTI systems in the presence of white noise is investigated. The single-pulse signal-to-noise ratios at the output of the filters are presented for the case of large system bandwidth. Generalized results, calculated as a function of the system's bandwidth, are given. Finally, those results are compared to the optimum signal-to-noise acheivable by a matched filter.     

Generalized Response of a Linear FM Pulse Compression Matched Filter

A pulse compression matched filter is analyzed so that the response may be computed when the pulse width, FM rate, and center frequency simultaneously differ from design conditions. Unilateral and bilateral time domain amplitude weighting for sidelobe reduction is included. A general cross-ambiguity function is defined to include these effects and some basic computed results are presented for the peak envelope response with various degrees of Hamming weighting. Computer evaluation of this cross-ambiguity function allows one to choose a combination of mismatches for signal design trade-off between resolution and detection performance. Since no restrictions are placed upon the mismatch parameters, this analysis may also be used to evaluate the filter discrimination against various interfering signals.     

Filtering of moving targets using SBIR sequential frames

In this paper three-dimensional (3-D) finite-impulse response (FIR) filters are proposed for moving target detection and tracking from multiframe space-based infrared (SBIR) data. An optimal, in the l_p sense, 3-D FIR filter design technique is proposed which is suitable for the above application. This technique is the first 3-D FIR design of its kind presented in the open literature. Directional, matched, and adaptive 3-D filtering techniques are proposed. Prior to the filtering, clutter mean estimation and mean subtraction are required. Real time implementation of directional and/or matched filters for processing maneuvering targets is discussed and filter design methods are proposed. Finally, performance comparisons of the proposed and other available 3-D FIR and infinite-impulse response (IIR) filters, on real SBIR data, are presented in which the advantages of the proposed 3-D filters are shown     

Interrelationship of single-filter and multiple-model adaptivealgorithms

An equivalent filter bank structure for multiple model adaptive estimation (MMAE) is developed that uses the residual and state estimates from a single Kalman filter and linear transforms to produce equivalent residuals of a complete Kalman filter bank. The linear transforms, which are a function of the differences between the system models used by the various Kalman filters, are developed for modeling differences in the system input matrix, the output matrix, and the state transition matrix. The computational cost of this new structure is compared with the cost of the standard Kalman filter bank (SKFB) for each of these modeling differences. This structure is quite similar to the generalized likelihood ratio (GLR) structure, where the linear transforms can be used to compute the matched filters used in the GLR approach. This approach produces the best matched filters in the sense that they truly represent the time history of the residuals caused by a physically motivated failure model     

Homodyned-K fluctuation model

The statistics of the matched filter output are examined for the case of a nonfluctuating target in K-type clutter. The resulting "homodyned-K" backscatter model is a generalization of the classic Rayleigh-Rice model and has applications in radar, sonar, optics, and ultrasonic medical imaging. This paper presents exact and asymptotic equations for the probability density function, survival function, moments, and parameter estimation. The exact equations are expressed as infinite series. Also provided are proof of the series' convergence and guidance for their numerical computation. Examples illustrating the model's application to radar and sonar are presented using electromagnetic and acoustic backscatter data     

Detection of chi-square fluctuating targets in arbitrary clutter

In a recent paper, general expressions were derived for the density and cumulative probability functions of the amplitude of a linear matched-filter output given a nonfluctuating target in a clutter-limited environment. These expressions were based on the clutter amplitude density function. The results are extended to calculate the cumulative probability function of the output of a linear matched filter used to detect a chi-square fluctuating target in a clutter-limited environment. The resulting method is applied to a common radar clutter model, and experimental sonar data.     

Optimum Mismatched Filters for Sidelobe Suppression

This paper discusses the application of least-mean-squares approximate inverse filtering techniques to radar range sidelobe reduction. The method is illustrated by application to the 13-element Barker code. The performance of the least-mean-square inverse filter is compared with the matched filter and with the simplified sidelobereducing filters of Rihaczek and Golden. A filter which completely suppresses the range sidelobes of a 13-element Barker sequence is only 0.2 dB worse than a matched filter in noise.     

Adaptive detection using low rank approximation to a data matrix

Using an accurate formula for the error in approximating a low rank component, we calculate the performance of adaptive detection based on reduced-rank nulling. In this principal component inverse (PCI) method, one temporarily regards the interference as a strong signal to be enhanced. The resulting estimate of the interference waveform is subtracted from the observed data, and matched filtering is used to detect signal components in the residual waveform. We also present a generalized likelihood-ratio test (GLRT) for adaptively detecting a low rank signal in the presence of low rank interference. This approach leads to a test which is closely related to the PCI method and extends the PCI method to the case where strong signal components are present in the data. A major accomplishment of the work is our calculation of the statistics of the output of the matched filter for the case in which interference cancellation and signal detection are carried out on the same observed data matrix. That is, no separate data is used for adaptation. Examples are presented using both simulated data and real, active-sonar reverberation data from the ARSRP, the Acoustic Reverberation Special Research Program of the Office of Naval Research     

Signal Resolution via Digital Inverse Filtering

Research in numerous areas is directed toward the resolution of multiple overlapping signals in a noisy environment. These areas include radar, sonar, speech, seismology, and electrophysiology. Sometimes matched filters are used; other times inverse filters are employed. This paper discusses one approach to the analysis of the resolution of inverse filters. Our method is to compromise the trade-off between signal resolution and the output signal-to-noise ratio (SNR). A performance measure for the inverse or deconvolution filter is defined as a quantity proportional to the harmonic mean of the resolution and the SNR. An optimum output pulse duration is obtained using this criterion, where the pulse shape has been previously selected and the input signal waveform is known. In addition, upper and lower bounds for the output pulse duration are presented. Graphs are given which allow the designer to select the optimum inverse filter output pulse duration for a desired signal resolution and an estimated SNR.     

Detection in correlated Gaussian plus impulsive noise

A detector which is designed to operate in a correlated Gaussian-plus-impulsive-noise environment is presented. The detector whitens the data robustly and then uses a two-sided threshold test to determine the presence of impulsive samples. The impulsive samples are discarded, and the remaining samples are used to detect the presence or absence of a signal using a matched filter. An approximate analysis is presented, and simulations are used to demonstrate the effectiveness of this approach     

Detection of Known Signals in Nonstationary Noise

The basic design of a nonlinear, time-invariant filter is postulated for detecting signal pulses of known shape imbedded in nonstationary noise. The noise is a sample function of a Gaussian random process whose statistics are approximately constant during the length of a signal pulse. The parameters of the filter are optimized to maximize the output signal-to-noise ratio (SNR). The resulting nonlinear filter has the interesting property of approximating the performance of an adaptive filter in that it weights each frequency band of each input pulse by a factor that depends on the instantaneous noise power spectrum present at that time. The SNR at the output of the nonlinear filter is compared to that at the output of a matched filter. The relative performance of the nonlinear system is good when the signal pulses have large time-bandwidth products and the instantaneous noise power spectrum is colored in the signal pass band.     

Effect of Particular Gain Changes upon LFM Sidelobes

The envelope variation of an LFM waveform due to transmitter droop or receiver STC tends to cause range sidelobes. A parametric analysis of the magnitude of the sidelobes has been performed. It is shown that the sidelobes can be quite high at the matched filter output, but are low at the output of the sidelobe reduction filter. 40-dB sidelobes can be achieved even with a 4-dB envelope droop. It is shown that these results are consistent with conventional paired-echo theory. Similar results are shown to hold for droop variations of the filter transfer function.     

Application of velocity filtering to optical-flow passive ranging

The performance of the velocity filtering method as applied to optical-flow passive ranging under real-world conditions is evaluated. The theory of the 3-D Fourier transform as applied to constant-speed moving points is reviewed, and the space-domain shift-and-add algorithm is derived from the general 3-D matched filtering formulation. The constant-speed algorithm is then modified to fit the actual speed encountered in the optical flow application, and the passband of that filter is found in terms of depth (sensor/object distance) so as to cover any given range of depths. Two algorithmic solutions for the problems associated with pixel interpolation and object expansion are developed, and experimental results are presented     

Compact AM Signals for Maximum Clutter Rejection

The transmitted signal is assumed to consist of a close succession of rectangular pulses of equal width. A matched filter scheme is employed and a theory is developed for a computer-aided optimization of the envelope of monotone compact signals for maximum rejection of dense clutter of any given distribution in range. Specific results are presented and indeterminate cases are discussed.