Research On Detection Algorithms Of Range-spread Targets

In recent years, with the increasing of radar range resolution, the target echo occupies several range cells, and the area of each range cell decreases obviously. Therefore, the traditional detection theory based on point-target and large-area clutter is no longer available. Therefore, the main problem of morden radar detection is detecting the range-spread targets.In this thesis, the detection algorithms of range-spread targets in Gaussian/compound Gaussian clutter have been investigated. The main results are as follows:1. Aiming at the problem of detecting range-spread targets in traditional compound Gaussian clutter, we propose the optimum detection strategies of the range-spread targets in K, Weibull and G0 clutter respectively. Comparing theses optimum detectors with the two-step (TS) GLRT detector, we conclude that the speackle covariance matrix estimator influence the performance of the TS GLRT detector a lot.2. Aiming at the problem of estimating the speckle covariance matrix, we proved that the the iterative MLE is optimum estimator of the covariance matrix.3. Aiming at the problem of detecting range-spread targets in compound Gaussian clutter without secondary data, we develop an AR-based detector named ARGLRT-CG detector, which has better detection performance (about 3 dB) than the known ARGLRT-HTG detector.4. Aiming at the problem that the ARGLRT-CG detector suffers performance loss when the coherent samples are few, we improve the ARGLRT-CG detector and propose another IARGLRT-CG detector, which has better performance (about 2~3 dB) than the ARGLRT-CG detector.Moreover, aiming at the problem of detecting range-spread targets in Gaussian clutter, we propose a Rao-based one-step (OS) AR-OS detector, which can be realized easier and has better rejection performance than the known AG-OS detector.The above detection algorithms have been proved by the simulation results, where the parameters have been estimated from the well-known IPIX radar clutter.