Research On Radar Target Detection Method Under Ground Clutter Background

Radar works in complex environments where it is usually interfered by all kinds of clutter.Therefore,how to suppress clutter and detect targets in complex background is an important research direction in radar field.With the development of radar automatic detection and recognition system,CFAR detection technology has become an important research direction in radar field.The clutter map has a good detection performance in checking target with slow speed or tangential radar.This paper focuses on the technology of clutter map and establishes the clutter map in each channel of the Doppler domain.By adding the Swerling target with slow speed in the actual scene,the results verify the better performance of clutter map CFAR detector for the slowing target compared with classic CFAR detector.Based on the shortcomings of the classic CFAR(i.e.,firstly,it is not to be ensured that the unit to be checked is independent from the reference unit.secondly,it is not guarantee that there is no target in the reference unit),this paper proposes a sample sharing method for similar clutter and overcomes the shortcomings of the classic CFAR and verifies its detection performance through simulation experiments.Firstly,the signal model,noise model and clutter statistical model are analyzed.The signal model is focused on the Swerling model and the simulation conditions of various Swerling models are introduced.Due to the noise model,we focus on the thermal noise model caused by radar receiver.The clutter model is mainly introduced several commonly used models of ground and see clutter and the adapt environment of various models are introduced.Secondly the necessity of CFAR detection is introduced by classic threshold detection.According to the Rayleigh clutter model,we infers the expressions of the false-alarm probability and detection probability of the unit average class CFAR,the double parameter class CFAR and the clutter map detector in the abstract.The clutter environment suitable for each detector and the merits and demerits of each detector are summarized.It also introduces an engineering method of clutter map CFAR detector.we also proved that this method has a good detection performance for slow moving target.Finally,based on the phenomenon that the two assumption of classical CFAR detection algorithm are dissatisfied,a similar sample sharing method is proposed.Through the cluster of the amplitude characteristics and Doppler characteristics are separately extracted from the actual measured data,we obtain the similar sample based on the amplitude characteristics and Doppler characteristics.For clustering results,we found that clustering results based on Doppler feature better than results based on amplitude feature.So the experimental part is based on Doppler feature to cluster and detect.It is found that experiment that clustering of mahalanobis distance or GMM through double parameter CFAR detector has the best performance.