Key radar signal fast recognition method based on clustering and time-series correlation

Aiming at the pertinence and ineffectiveness of the traditional radar signal recognition method in identifying key targets, a real-time radar signal recognition method based on clustering and time-series correlation is proposed. Firstly, pulse description words of detected signals are sorted based on density-based spatial clustering of the application with noise (DBSCAN) algorithm. Then, the timing characteristics of the sorting pulse and the pulse repetition interval (PRI) parameters of the key target signal are used to generate the simulation signal. Finally, the cross-correlation function of the simulated signal is calculated, and the PRI parameter is judged to be matched based on the degree of correlation. Simulation results show that the proposed method significantly improves the identification time of key target signals, can deal with the complex signal environment with noise interference and overlapping signals, and is not sensitive to the loss of local pulse parameters. © 2020, Editorial Office of Systems Engineering and Electronics. All right reserved.