Adaptive Target Detection of High-range-resolution Radar in Non-Gaussian Clutter

摘要

Adapti ve target detection of high-range-resolution radar in spherically invariant random vector clutter is investigated for a possibly singular estimated covariance matrix. Under an appropriate constraint on the unknown parameter space, the maximum likelihood estimates are obtained. By utilizing the modified recursive estimator for possibly singular covariance matrix estimation, the detector of adaptive restricted generalized likelihood ratio is devised for non-Gaussian scenarios. The results show that the proposed detector outperforms the existing Kelly binary integrator. As the number of equivalent scatterers increases, the detection performance improves whereas the performance gain decreases. Moreover, in the range of high values of detection probability, the detection performance decreases as the clutter spikiness increases, while the performance increases with increasing clutter spikes in the range of low values of detection probability.