A Study Of Ground Radar Dim Target Detection

Nowadays, radar is faced with various dangers and challenge and dim target detection is an important task in the field of radar signal processing. In order to detect the dim targets, ways such as long time coherent accumulation, track-before-detect, are came up with to improve the SNR. The two methods about dim radar detection are discussed in the article.In order to solve range migration and Doppler walk caused by target motion during long time coherent integration period, a novel algorithm based on frequency domain correction is proposed for moving target with constant acceleration which is implemented by multiplying echoes after range compression in frequency domain by the acceleration correction factor and the velocity correction factor, respectively. The acceleration is estimated by Radon-Ambiguity transformation(RAT), and the velocity is estimated by searching the peak of the chosen cost function. The results of simulation prove it can not only correct range migration and Doppler walk effectively, but also reduce the complexity. In order to overcome the reasonable choice of the importance proposal distribution and the sample impoverishment after resampling, an improved particle filter algorithm based on evolution fission is presented. The algorithm uses UKF to generate the importance proposal distribution and updates the samples based on evolution fission, which could increase the diversity of the samples.The results of simulation prove it has high tracking accuracy and tracking performance compared to the typical particle filter. Finally, several common detectors’ performance is discussed in different conditions.