Study On Constant False Alarm Rate Detection Algorithms With Coherent Radar

With the development of coherent radar technology, constant false alarm (CFAR) detection with coherent radar has become an interesting issue among world radrar community in the two decades. Gaussian clutter model has been generally used as that arises from the central limit theorem. But it is unable to describe the statistics of clutter for high-resolution radar or at low grazing angles. Experimental data indicates that the non-Gaussian clutter can be modeled as a compound-Gaussian process. This dissertation aims at the study of CFAR detection in gaussian and compound-gaussian clutter with coherent radar. The main work can be summarized as follows:1. The problem of polarimetric adaptive detection with coherent radar in a gaussian background is addressed. Previously proposed single channel and double channel adaptive matched filter are extended to the case of a general number of channels. Polarimetric adaptive matched filter (PAMF) with coherent radar is presented. This exploits the polarimetric characteristics of the received radar echoes to improve the detection performance. A model-based polarimetric adaptive matched filter (MPAMF) is presented that has a lower estimation loss. It is shown that the proposed detection algorithms have CFAR property with respect to the covariance matrix of the clutter. Resorting to simulation,the performances of both algorithms are analyzed in different polarimetric cases.2. Polarimetric adaptive detection of targets with coherent radar in partially-homogeneous Gaussian clutter background is addressed. Based on generalized likelihood ratio test (GLRT), polarimetric adaptive subspace detector (PASD) is derived. The theoretical expressions of probability of false alarm and probability of detection for PASD are derived. The proposed detection algorithm ensures CFAR in homogeneous and partially-homogeneous clutter background. Resorting to simulation, the performance of PASD is analyzed in partially-homogeneous clutter in different polarimetric cases.3. Polarimetric adaptive detection of targets in compound-Gaussian clutter with coherent radar is addressed. An adaptive polarimetric detection algorithm is presented based upon GLRT. The proposed algorithm has CFAR property with respect to the texture statistical characterization. A sensitivity analysis shows that the probability of false alarm is only slightly affected by variations in the clutter correlation properties. Resorting to simulation, the performance of the algorithm is analyzed in different cases. The proposed algorithm has a close performance as previous proposed algorithm, but it has less complexity than the previously proposed one.4. Adaptive detection of distributed targets with the coherent radar in compound-Gaussian clutter background is addressed. By exploiting the persymmetry property of toeplitz covariance matrix, we proposed a new method of covariance matrix estimation to implement CFAR detection algorithm based on GLRT. Without needing the symmetric clutter power spectral density, the proposed detection algorithm ensures CFAR the with respect to both the clutter covariance matrix as well as the statistics of the texture. It is shown that the proposed one outperforms the previously proposed one.5. Multiband adaptive detection of targets in compound-Gaussian clutter background is addressed. A multiband adaptive detector is derived for the coherent radar detection in a compound-gaussian clutter background. This scheme is an extension to multiband case of the adaptive normalized matched filter. The proposed multiband version of the algorithm shows an advantage with respect to the original single-band algorithm. For adaptive implementation, it requires a much smaller area of homogeneous clutter echoes to estimate the covariance matrix of the clutter. It ensures the CFAR property with respect to both the clutter covariance matrix as well as the statistics of the texture.By exploiting the persymmetry property of toeplitz matrix, a new multiband coherent radar adaptive CFAR detection algorithm is presented. The proposed algorithm ensures CFAR with respect to both the clutter texture as well as the clutter covariance matrix, and does not need the symmetric clutter power spectral density. The proposed algorithm outperforms the previously proposed one.