Convergence properties of Gram-Schmidt and SMI adaptive algorithms.II

For pt.I see ibid., vol.26, no.1, p.44-56, Jan. 1990. Theorems and relationships associated with the convergence rate of the Gram-Schmidt (GS) and sampled matrix inversion (SMI) algorithms are presented. Two forms of the GS canceler are discussed: concurrent block processing and sliding window processing. It is shown (as has been stated by other researchers) that the concurrent block processed GS canceler converges rapidly to its optimal signal-to-noise ratio. However, it is also shown that the result is deceptive in that the output residue samples may be highly correlated, which would significantly degrade postdetection processing. It is demonstrated that a specific form of a sliding window GS canceler has the same convergence properties as the concurrent block processed GS canceler     

Adaptive array transient sidelobe levels and remedies

The transient sidelobe level of a sidelobe canceler (SLC) is a function of the external noise environment, the number of adaptive auxiliary antennas, the adaptive algorithm used, auxiliary antenna gain margins, and the number of samples used to calculate the adaptive weights. An analytical result for the adaptive sidelobe level is formulated for the case when the adaptive algorithm is the open-loop, sampled matrix inversion (SMI) algorithm. The result is independent of whether concurrent or nonconcurrent data processing is used in the SMI algorithm's implementation. It is shown that the transient sidelobe level is eigenvalue dependent and increases proportionally to the gain margin of the auxiliary antenna elements with respect to the quiescent main antenna sidelobe level. Techniques that reduce this transient sidelobe level are discussed, and it is theoretically shown that injection independent noise into the auxiliary channels significantly reduces the transient sidelobe level. It is demonstrated that using this same technique reduces the SMI noise power residue settling time     

Convergence bounds of an SMI/Gram-Schmidt canceler in colored noise

The performance of the sampled matrix inversion (SMI) adaptive algorithm in colored noise is investigated using the Gram-Schmidt (GS) canceler as an analysis tool. Lower and upper bounds of average convergence are derived, indicating that average convergence slows as the input time samples become correlated. When the input samples are uncorrelated, the fastest SMI algorithm convergence occurs. When the input samples are correlated then the convergence bounds depend on the number of channels N, the number of samples per channels K , and the eigenvalues associated with K×K correlation matrix of the samples in a given channel. This matrix is assumed identical for all channels     

Reiterative median cascaded canceler for robust adaptive array processing

A new robust adaptive processor based on reiterative application of the median cascaded canceler (MCC) is presented and called the reiterative median cascaded canceler (RMCC). It is shown that the RMCC processor is a robust replacement for the sample matrix inversion (SMI) adaptive processor and for its equivalent implementations. The MCC, though a robust adaptive processor, has a convergence rate that is dependent on the rank of the input interference-plus-noise covariance matrix for a given number of adaptive degrees of freedom (DOF), N. In contrast, the RMCC, using identical training data as the MCC, exhibits the highly desirable combination of: 1) convergence-robustness to outliers/targets in adaptive weight training data, like the MCC, and 2) fast convergence performance that is independent of the input interference-plus-noise covariance matrix, unlike the MCC. For a number of representative examples, the RMCC is shown to converge using ~ 2.8N samples for any interference rank value as compared with ~ 2N samples for the SMI algorithm. However, the SMI algorithm requires considerably more samples to converge in the presence of outliers/targets, whereas the RMCC does not. Both simulated data as well as measured airborne radar data from the multichannel airborne radar measurements (MCARM) space-time adaptive processing (STAP) database are used to illustrate performance improvements over SMI methods.     

An MTI Technique for Processing Radar Measurements of a Waking Reentry Body

A processing technique based on pulse-cancellation techniques familiar in moving target indicator (MTI) radar is proposed for separating (in Doppler) echoes of a reentry body traveling at hypersonic velocities from those of its lower velocity turbulent wake appearing in the same range cell. The cancellation technique is implemented by forming the sum of the products of binomial weighting coefficients of alternating sign with the complex echoes of a small number of closely spaced transmitted coherent pulses; thereby, in effect, synthesizing a digital canceler. The ability of the two-and three-pulse canceler to estimate body RCS in the presence of attached wake is demonstrated by employing coherent burst data collected by the AMRAD radar for a mission flown at the White Sands Missile Range. Estimates of body RCS obtained from the two-and three-pulse canceler compare favorably to the corresponding estimates obtained from a 30-pulse Doppler periodogram for this mission. Expressions for both the achievable wake rejection ratio and the mean and standard deviation of the body power estimate of the N-pulse canceler are derived as a function of the wake parameters, assuming Gaussian wake statistics.     

Adaptive canceler and pulse compressor interactions

Performance results for the sidelobe level of a compressed pulse that has been preprocessed through an adaptive canceler are obtained. The adaptive canceler is implemented using the sampled matrix inversion algorithm. Because of finite sampling, the quiescent compressed pulse sidelobe levels are degraded due to the preprocessing of the main channel input data stream (the uncompressed pulse) through an adaptive canceler. It is shown that if N is the number of input canceler channels (main and auxiliaries) and K is the number of independent samples per channel, then K/N can be significantly greater than one in order to retain sidelobes that are close to the original quiescent sidelobe level (with no adaptive canceler). Also it is shown that the maximum level of degradation is independent of whether pulse compression occurs before or after the adaptive canceler if the uncompressed pulse is completely contained within the K samples that are used to calculate the canceler weights. This same analysis can be used to predict the canceler noise power level that is induced by having the desired signal present in the canceler weight calculation     

Implementation and convergence considerations of a linearlyconstrained adaptive array

O.L. Frost (1972) introduced a linearly constrained optimization algorithm that allows certain main beam properties to be preserved while good cancellation is attained. An open-loop implementation of this algorithm is developed. This implementation is shown to be equivalent to the technique developed by C.W. Jim (1977), L.J. Griffiths and C.W. Jin (1982), and K.M. Buckley and L.J. Griffiths (1982) whereby the constrained problem is reduced to an unconstrained problem. Analytical results are presented for the convergence rate when the sampled matrix inversion (SMI) or Gram-Schmidt (GS) algorithm are employed. It has been previously shown that the steady-state solution for the optimal weights is identical for both constrained and reduced unconstrained problems. It is shown that if the SMI or GS algorithm is employed, then the transient weighting vector solution for the constrained problem is identical to the equivalent transient weight vector solution for the reduced unconstrained implementation     

Performance comparisons of modified SMI and GLR algorithms

Adaptive filtering for signal detection in colored interference of unknown statistics is addressed. The detection performance of a modified version of the well-known sample matrix inversion (SMI) algorithm, called the modified SMI (MSMI), is compared with that of the generalized likelihood ratio (GLR) algorithm in colored Gaussian interference. The performance sensitivity of the MSMI and GLR in colored Weibull and log-normal interference is studied via simulation. It is found that there is almost no need to use the more complicated GLR algorithm in Gaussian interference, while in Weibull or log-normal interference the GLR should be preferable to the MSMI     

Trimmed Generalized Sign and Modified Median Detectors for Multiple Target Situations

The trimmed generalized sign (TGS) nonparametric detector is introduced. The TGS and the modified median detector (MMD) are considered in situations when more than one target is present. Their performance is obtained through Monte Carlo simulations and compared with that of the generalized sign (GS) detector when detecting nonfluctuating signal in Gaussian noise.     

A sampling-based approach to wideband interference cancellation

Classical adaptive array schemes which use only complex spatial weights are inherently narrowband and consequently perform poorly when attempting to suppress wideband interference. The common solution to this problem is the use of tapped delay line filters in each spatial channel to facilitate space-time adaptive processing (STAP). The higher performance provided by the STAP architecture comes at the cost of a considerable increase in complexity. This paper presents a simpler technique based on programmable time adjustable sampling (TAS) that provides a limited number of wideband degrees of freedom. Two TAS methods are introduced: TAS-sidelobe canceler (TAS-SLC) is based on the sidelobe canceler, while TAS-minimum variance beamformer (TAS-MVB) is derived from the minimum variance beamformer. TAS is implemented by adjusting the sampling instant at selected array channels. TAS-SLC consists of controlling the sampling in the main channel of the sidelobe canceler With TAS-MVB array complex weights are substituted with TAS time delays. The performance of TAS methods with wideband interference is compared to the conventional sidelobe canceler and minimum variance beamformers. It is shown that TAS-SLC provides better performance than the sidelobe canceler, while TAS-MVB outperforms the minimum variance beamformer     

Steered Beam and LMS Interference Canceler Comparison

The performance of two kinds of interference cancelers is compared, namely, a constrained steered beam interference canceler (CSBIC) and a least mean square interference canceler (LMSIC). For simplicity, this is done for a two-element array. In our comparison we use the array output desired signal-to-interference power ratio (SIR) and the desired signal-to-noise ratio (SNR). These power ratios are obtained for the CSBIC and LMSIC under two sets of conditions: 1) The error in the assumed angle of incidence for the CSBIC is small, and the LMSIC operates in a codetracking mode. 2) The error in the assumed angle of incidence for the CSBIC is large, and the LMSIC operates in a code-acquisition mode. Comparison of the corresponding power ratios obtained under these two sets of conditions then establishes the condition under which it is more desirable to use a CBSIC as compared with an LMSIC.     

Modified Savage and Modified Rank Squared Nonparametric Detectors

A modified form of the basic Savage statistic is considered and the performance of a modified Savage (MS) nonparametric detector using this modified statistic is derived. Also, a detector using a modified rank squared statistic (MRS) is introduced. The asymptotic relative efficiency (ARE) of the detectors is determined for chisquare, Rician, and log-normal signal fluctuations when the background noise is assumed Gaussian. The ARE performance of the generalized sign (GS) and Mann-Whitney (MW) detectors is also determined for these families of fluctuations. The ARE performance of the various detectors is then compared, and the results of a computer simulation are presented in which, for a finite number of samples, the performance of the modified detectors is compared with the performance of the GS and MW detectors. It is shown that when using a large number of reference noise samples, the ARE of the GS and MW detectors, the MRS and RS detectors, and the MS and Savage detectors are 0.75, 0.868, and 1, respectively. It is also shown that when using a finite number of reference noise samples the MS and MRS detectors can give a superior performance to that obtained with the MW detector, and that this is particularly true in the cases in which the degree of signal fluctuation is high.     

An Adaptive MTI for Weather Clutter Suppression

A digital realization of an adaptive clutter-locking loop is presented. The purpose of the loop is to estimate the mean Doppler frequency of the clutter. The clutter spectrum is then shifted toward the zero Doppler by this estimate. A fixed moving target indicator (MTI) canceler following the loop suppresses the shifted clutter. Experimental simulations illustrate the feasibility of the loop. Results indicate that the proposed canceler works significantly better than a fixed canceler, while not as well as the 10-pulse moving target detector (MTD) processor. However, the complexity of the MTD is significantly more than the relatively simple adaptive processor presented here.     

海杂波中广义符号恒虚警检测算法性能分析

分析了广义符号检测算法在仿真的高斯杂波背景和实测海杂波背景下,对2种目标（Sweding0型和Swerling II型）的检测性能,以及对实际渔船目标的检测性能。研究表明,随着脉冲数、参考单元数和信杂比的提高,该检测算法的检测性能有所提高;在低信杂比条件下,GS检测算法对SwedingII型目标的检测性能优于对Sweding0型目标的检测性能,在高信杂比的条件下,对Swerling 0型目标的检测性能优于对Swerling II型目标的检测性能。     

Detection Loss of the Sample Matrix Inversion Technique

The sample matrix inversion (SMI) technique is used for Doppler and/or array processing. Previous analysis of the technique has been in terms of signal-to-interference plus noise ratio (SINR). For Gaussian statistics, this performance measure gives the same loss values as does a probability of detection analysis for linear-time invariant systems. It is often somewhat less valid for nonlinear or time variant systems. As SMI is a nonlinear technique, a probability of detection analysis has been performed. It is shown that the detection loss is larger than that computed by the SINR measure. It is also shown that though the loss predicted by the SINR measure only depends upon the number of measurements used to estimate the covariance matrix, the detection loss depends upon the false alarm probability and the number of adaptable elements in addition to the number of measurements.     

The effect of I,Q mismatch errors on adaptive cancellation

The effects of in-phase (I) and quadrature-phase (Q) amplitude errors and low-pass-filter (LPF) errors on adaptive cancellers are investigated. I,Q errors occur because of errors in the synthesis process of the mixers and LPFs designed to be identical for each input channel. These I,Q errors among the channels result in cancellation degradation. Tapped delay line transversal filters have been proposed as a way to compensate for these errors and thus improve cancellation performance. However, it is shown that if there is any LPF mismatch, then transversal filtering has a small effect on improving canceler performance. The use of individual I,Q adaptive transversal filter weighting is suggested as a means of completely eliminating the phase amplitude errors, and making the canceler performance responsive to transversal filter compensation     

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     

Design and Evaluation of an Adaptive MTI Filter

The design and evaluation of an adaptive moving target indicator (MTI) filter, the adaptive canceler for extended clutter (ACEC) is dealt with, taking into consideration adaptivity to clutter mean Doppler frequency. This consideration is one of the most important operational requirements in adaptive MTI's and permits a relatively simple hardware implementation as compared to more general optimization and adaptivity criteria (briefly described). The ACEC's algorithm compensates in real time for the clutter mean Doppler frequency. Performances have been obtained by digital computer simulation in various operational conditions.     

A new CFAR sidelobe canceler algorithm for radar

Signal or target detection is sometimes complicated by the presence of strong interference. When this interference occurs mainly in the sidelobes of the antenna pattern, a solution to this problem is realized through a sidelobe canceler (SLC) implementation. Since the false-alarm probability is a system parameter of special importance in radar, an interference-canceling technique for radar application should maintain the false-alarm probability constant over a wide range of incident interference power. With the requirements of sidelobe interference cancellation and constant false alarm rate (CFAR), a new algorithm for radar detection in the presence of sidelobe interference is developed from the generalized likelihood ratio test of Neyman-Pearson. In this development, the received interference is modeled as a nonstationary but slowly varying Gaussian random process. Cancellation of the sidelobe interference is based upon a `synchronous' estimate of the spatial covariance of the interference for the range gate being tested. This algorithm provides a fixed false-alarm rate and a fixed threshold which depend only upon the parameters of the algorithm     

A Detection Algorithm for Optical Targets in Clutter

There is active interest in the development of algorithms for detecting weak stationary optical and IR targets in a heavy opticalclutter background. Often only poor detectability of low signal-to-noise ratio (SNR) targets is achieved when the direct correlation method is used. In many cases, this is partly obviated by using detection with correlated reference scenes [1, 2].This paper uses the experimentally justified assumption that most optical clutter can be modeled as a whitened Gaussian randomprocess with a rapidly space-varying mean and a more slowlyvarying covariance [2]. With this assumption, a new constant falsealarm rate (CFAR) detector is developed as an application of the classical generalized maximum likelihood ratio test of Neyman and Pearson. The final CFAR test is a dimensionless ratio. This test exhibits the desirable property that its probability of a false alarm(PFA) is independent of the covariance matrix of the actual noiseencountered. When the underlying noise processes are complex intime, similar considerations can yield a sidelobe canceler CFARdetection criterion for radar and communications. Performance analyses based on the probability of detection (PD)versus signal-to-noise ratio for several given fixed false alarm probabilities are presented. Finally these performance curves are validated by computer simulations of the detection process which use real image data with artificially implanted signals.