Abstract
In this paper, we develop theories and an algorithm for the detection and estimation of high-velocity targets using pulse-Doppler radars. First, we prove that the high velocities of targets lead to range and velocity migrations proportional to the velocities, thus leading to block-sparse structure of the range-velocity spectrum. Then, based on this theory, we formulate the detection and estimation task as a block-sparse recovery problem and propose an algorithm named pulse-Doppler total variation regularized sparse Bayesian learning (PD-TVSBL) to solve this problem, thus recovering the range-velocity spectrum of the targets. Simulation results show that the proposed method achieves a gain of 11 dB in terms of signal-to-noise ratio (SNR) for accurate detection compared to the conventional iterative soft thresholding (IST) method. Moreover, the performance of the proposed method approaches that of a genie-aided method, whose performance is the theoretical optimum of block-sparse modeling, in the medium to high SNR regime. This paper provides a solid solution for accurate detection and estimation of high-velocity targets for pulse-Doppler radars.
Funding Information
  • National Natural Science Foundation of China (62388102)
  • National Natural Science Foundation of China (62571297)
  • Shenzhen Key Laboratory of Ubiquitous Data Enabling (ZDSYS20220527171406015)
  • Shenzhen Science and Technology Program (JCYJ20240813112301003)
  • Tsinghua Shenzhen International Graduate School-Shenzhen Pengrui Endowed Professorship Scheme of Shenzhen Pengrui Foundation

This publication has 36 references indexed in Scilit: