Distributed Target Detection in Compound-Gaussian Clutter Under Steering Vector Uncertainty

Track-before-detect strategies for range distributed target detection in compound-Gaussian clutter

This paper provides a historical and tutorial overview of coherent radar target detection in compound-Gaussian clutter and offers some new perspectives and avenues of research in this challenging area. It begins with a brief introduction that motivates the need to develop statistical models of non-Gaussian clutter and then reviews some of the physical ideas that led to modeling multivariate radar clutter statistics by the compound-Gaussian model. With this starting point, the paper then reviews a series of ideas that have been developed to describe clairvoyant detectors in such clutter. The term “clairvoyant” refers to the assumption that the properties of the clutter are assumed to be known. In a practical scenario, this assumption does not hold and adaptive techniques are needed to estimate clutter properties and implement the detector. Such techniques are guided however by the appropriate clairvoyant detector structures and hence it is proper to start by studying these detectors. As part of this review, the paper offers new ways of looking at this problem that suggest new research topics. This review is limited to the problem of clairvoyant detection in which the relevant properties of the clutter are assumed to be known. Adaptive detection in compound- Gaussian clutter will be the topic of a subsequent tutorial that the authors are preparing.

Detection of weak moving targets with a low signal-to-clutter ratio (SCR) in heavy-tailed non-Gaussian sea clutter is a challenge problem in radar signal processing. In this study, an easily implementable suboptimum detector in compound-Gaussian (CG) clutter is proposed based on suprathreshold stochastic resonance (SSR). The structure of the locally optimum coherent detector for weak target detection in CG clutter is first analysed and the authors proved that it can be implemented using a pre-processing system followed by the adaptive matched filter (AMF) used in Gaussian backgrounds. A suboptimum pre-processing system is designed using the SSR system consisting of a parallel array of quantisers, which overcomes the high computation cost caused by the calculation of the data-dependent threshold for optimum detectors. Its input–output characteristic can be adjusted by adding different quantiser noise at the input of the system, making it suitable to be used for different distributions of input clutter. The gain of the SSR pre-processing system is introduced and the parameter determination for quantiser noise is analysed. Experiments on both simulated clutter and real sea clutter data in the Council for Scientific and Industrial Research's database and Intelligent PIXel Grimsby database have shown that the proposed detector based on SSR could achieve robust detection performance and perform better than the existing methods.

We focus on the problem of detecting a distributed target in compound-Gaussian clutter, where the texture is a random variable with Gamma distribution. The Rao and Wald tests are devised by using the two-step method: in the first step, the Rao and Wald tests are designed by assuming that the texture and covariance matrix structure are known; in the second step, the maximum a posterior probability estimate of clutter texture and the fixed point estimate of covariance matrix structure are used to replace the known texture and covariance matrix in the tests derived in the first step, respectively. The effectiveness of the proposed detectors is verified by using simulated and real data.

Adaptive range-spread maneuvering target detection in compound-Gaussian clutter

Eigenvalue-based distributed target detection in compound-Gaussian clutter

Subspace-based distributed target detection in compound-Gaussian clutter

This paper considers moving target detection for multiple-input multiple-output (MIMO) radar in compound-Gaussian clutter. A new detector is devised according to a centralized processing scheme for MIMO radar system based on the generalized likelihood ratio test (GLRT) design criterion. Then, an adaptive version of the derived detector is investigated. The fixed point estimation (FPE) strategy is introduced to make the proposed detector fully adaptive. Finally, several numerical simulations of the derived detectors with typical parameters are obtained and discussed.

Multiple-input multiple-output (MIMO) radars with wideband signals are useful to discriminate target in clutter. With wideband MIMO radar, the problem of extended target detection in compound-Gaussian clutter is considered in this paper. By obtaining target echoes from each antenna, the detector based on generalized likelihood ratio test (GLRT) is developed. It shows that the shape parameter of clutter, Doppler mismatches and antenna configurations of MIMO radar heavily affect the performance of GLRT detector. Therefore, the detection performance is respectively analysed with different shape parameters and Doppler mismatches. By using maximum likelihood estimation (MLE), the detection curves with typical parameters are presented to illustrate that the proposed detector outperforms that obtained in Gaussian clutter. Then, the detection performance is discussed with different antenna configurations. Mathematical analysis and simulation results show that the detection performance reaches the best when the numbers of transmitters and receivers are equal, which provides reference for the antenna configuration of MIMO radar.

Adaptive two-step Bayesian MIMO detectors in compound-Gaussian clutter

This paper studies adaptive target detection in compound-Gaussian clutter for a colocated multi-input multi-output (MIMO) radar. A new one-step generalized likelihood ratio test (GLRT) is devised with all the involved unknown parameters estimated iteratively. Besides, a minor revision is made to the already existing two-step GLRT for this problem. Numerical examples show that both the two detectors are effective. For the considered simulation setup, the newly proposed detector can achieve higher detection rates than the revised one and competitors designed in Gaussian clutter scenarios.

Proper prior knowledge of target scattering centers (SCs) can help to obtain better detection performance of range-spread targets. However, target SCs are sensitive to the target's attitude relative to the radar and vary significantly among different targets. The existing approaches that employ predetermined prior knowledge may suffer performance degradation when the prior information does not match the practical scenarios. A possible way to circumvent this drawback is to estimate the SCs of different targets adaptively and check the presence of a target utilizing the range cells occupied by the most likely target SCs. For this reason, this article develops a generalized likelihood ratio test based on adaptive SCs estimation (ASCE-GLRT) for range-spread target detection in compound-Gaussian clutter. Under the assumption that the target SCs are sparse, we model the problem of SCs estimation as a sparse signal representation. Moreover, since the sparse assumption may not always be satisfied in practice, a modified sparsity regularization method is proposed to enhance the robustness of the estimation performance of targets with different scattering characteristics. A theoretical analysis shows that the proposed detector can achieve the constant false alarm rate (CFAR) property. The performance assessments conducted by numerical simulation and field tests confirm the effectiveness and robustness of the proposed detector.

In this paper, we consider polarimetric adaptive detection in compound Gaussian clutter whose covariance matrix (CM) has a Kronecker structure. We derive the Cramer–Rao bound of a Kronecker structured CM estimate and analyze the constant false alarm rate property of the adaptive subspace matched filter detector that uses Kronecker structured estimates. We provide a general expression for the average signal-to-clutter ratio loss (SCRL) as a function of the mean square error of the covariance estimate. The aforementioned expression is helpful in determining how many samples are required in order to achieve a desired average SCRL level in practical scenarios. Based on that expression, we show that the required sample size can be effectively reduced by exploiting the Kronecker structure of the clutter CM. We also derive the asymptotic detection performance of the adaptive subspace matched filter. The analysis of SCRL and detection performance can be extended to more general scenarios, especially when the maximum-likelihood estimate of the structured CM involves solving fixed point equations. Numerical simulations validate the merits of the proposed methods.

This paper proposes orthogonal Discrete Frequency Coding Space Time Waveforms (DFCSTW) for Multiple Input and Multiple Output (MIMO) radar detection in compound Gaussian clutter. The proposed orthogonal waveforms are designed considering the position and angle of the transmitting antenna when viewed from origin. These orthogonally optimized show good resolution in spikier clutter with Generalized Likelihood Ratio Test (GLRT) detector. The simulation results show that this waveform provides better detection performance in spikier Clutter.

This paper proposes orthogonal Discrete Frequency Coding Space Time Waveforms (DFCSTW) forMultiple Input and Multiple Output (MIMO) radar detection in compound Gaussian clutter.The proposed orthogonal waveforms are designed considering the position and angle of the transmitting antenna when viewed from origin.These orthogonally optimized show good resolution in spikier clutter with Generalized Likelihood Ratio Test (GLRT) detector.The simulation results show that this waveform provides better detection performance in spikier Clutter.