Research On Algorithms For LPI Radar Reconnaissance Signal Processing

It’s an important part of electronic intelligence to process the intercepted radar signal in the area of electronic warfare. According to the complex electromagnetic environment, the LPI radar developed rapidly. It’s the primary task to detect and analysis the intercepted radar signal in order to reconnoiter and interfere with the enemy LPI radar. As a result, it is of sovereign importance to research on the theory and algorithm for LPI radar signal processing. In this work, the techniques for typical LPI radar signal processing based on the previous research results are studied. The principal contributions are summarized as follows:First of all, the principles of LPI radar and factors which may affect inception probability are introduced, and the research status of LPI radar signal processing is also proposed.Then, the parameter estimation of LFM signal based on Fr FT is studied. Two interpolation methods of parameter estimation are presented to make up the “picket fence effect” due to the search step limitation and parameter discretization, which leads to the improvement of parameter estimation under low SNR conditions. What’s more, based on the presented methods, an iterative interpolation algorithm is further introduced, which results in perfect estimation result closed to CRLB.The third part is about the recognition methodology for three typical hybrid modulation signal, including LFM/BPSK hybrid modulation signal, FSK/BPSK signal and S-type NLFM signal. In the beginning of this part, the instantaneous frequencies of hybrid modulation signals and their characters are analyzed. Then, the fundamental of recognition threshold based on N-P criterion is discussed in detail, and the effect of threshold on the recognition probability is also involved.Fourth, the multiple parameter estimation for hybrid modulation signal is researched, including the previous three kinds of typical hybrid signals. The modulation mode separation is the base, and then the parameter estimation can be performed step by step as that of mono-modulation signal. For S-type NLFM signal, we mainly introduce the estimation of parameter bandwidth and middle frequency.Finally, an improved segment filter for wideband signal denoising is studied. The segment filter is achieved through a set of steps, including segment, frequency shift and windowing in frequency domain, and the main computation is FFT, which leads to easy realization in engineering.