Research On Constant False Alarm Rate Detection For MIMO Radar

Inspired by multiple-input multiple-output (MIMO) communications, fishler proposed the concept of MIMO radar. Currently, the researches on MIMO radar are focused on two main configurations: the first one is base on the phased-array radar with the distance between its arrays is half wavelength. It realizes the MIMO characteristic by utilizing waveform diversity; the second one is base on the multi-static radar. The space between the arrays is far enough in this type of MIMO radar, so it can get space diversity, thus can improve the performance of target detection and parameter estimate. This paper mainly researches the target detection of MIMO radar, including the following issues:Base on the cell-average constant false alarm rate detector, we proposal a double threshold Constant False Alarm Rate (CFAR) detector. Fist, we analyze the detection performance of the double threshold CFAR detector, and then compare it to the single threshold detector, and analyze this detector’s performance in multiple targets environment last. The simulation results show that: applying double threshold detection in MIMO radar, we can improve the constant false alarm rate detection performance, especially in multiple targets environments. Besides it has the advantage of simplicity, low data transmission rate and so on.In the non-Gaussian clutter environments, which are modeled as Spherically Invariant Random Vector (SIRV), we analyze the adaptive target detection problem on MIMO radar. After extending the generalized likelihood rate test (GLRT) to the MIMO radar detection, we derive the detector in SIRV clutter environments. In order to prove this new detector’s robustness, we compare it with the detector derived in the Gaussian clutter environments. And then researches the CFAR detection problem of MIMO radar in non-Gaussian clutter environment and analyses the target detection performance in K-distribute clutter environment. At last we make some simulations with Monte-Carlo method. The simulation results show that: the MIMO-GLTR detector maintains good detection performance in Gaussian and non-Gaussian clutter environment and performances obviously better than the optimization Gaussian detector.Based on Extreme Value Theory and Kolmogorov-Smirnov statistical test, this paper analyses the detection performance of MIMO radar. After introducing the extreme value theory and kolmogorov-smirnov statistical test, we use this theory to detect target in MIMO radar. The method proposed by this paper can estimate the detection threshold accurately under the environments that the distribution of the backgrounds is unknown and multiple targets are presented. Besides, it can accurately estimate the targets’number, thus improves the detection performance under multiple targets environment.In allusion to the waveform diversity in MIMO radar, we research the MIMO radar target detection problem base on space time waveform adaptive process, and then apply the generalized likelihood rate test to this problem. The parameters of the probability statistics distribute change after space time waveform adaptive process, and the generalized likelihood rate test can detect targets in the unknown Gaussian clutter environment. This paper adopts the generalized likelihood rate test detection algorithm and makes some theoretic analysis, then we checkout the detector’performance in the way of Monte-Carlo simulation.