A Dictionary Learning Method for High-resolution Detection Under Interference Background

This paper investigates the problem of high-resolution detection of target parameters by compressed sensing radar (CSR) in an interference background. The paper first assigns a distinct initial phase to each transmitted pulse to obtain the separability of the real target echo from the spoofed interference signal. And the suppression of the spoofed interference signal is realized by designing a measurement matrix to decode the added initial phase of each received pulse and compile the target echo. On this basis, the authors further propose a secondary reconstruction technique based on dictionary learning, which realizes high-resolution estimation of the target parameters through dynamic meshing. Compared with the traditional fixed-grid sparse-domain processing method, this method significantly improves the accuracy of target parameter estimation. The numerical results given in the paper show the superior effectiveness of the proposed deceptive interference suppression method, especially in high-resolution target detection. This suggests that the method is effective in mitigating the negative impact of repetitive interference on the detection performance of the CSR system, thereby accurately identifying and localizing targets in the presence of interference.