The Recognition Technology Of The Radar Signals' Intra-pulse Modulation

Radar signal classification and recognition is a method that separates the signals with different modulation modes according to the significant difference of parameter feature. It can lay the foundation for the following signal processing and provide the reliable information necessary for combatant commanders. A common method of radar signal recognition in the world is the technology of feature recognition by instantaneous frequency based on STFT and Wavelet Transform and this analysis only applies to small electromagnetic density and less complex electromagnetic signal environment with some prior knowledge. However, the radar signals processed by electronic reconnaissance system often complicated and the prior knowledge is difficult to obtain.How to make radar signal recognition, parameters estimation and reliability detection efficient is the key to the signal processing of the reconnaissance system.In this paper, the radar signals are coarse classified on account of the bandwidth characteristic and precise classified based on the instantaneous frequency and high-order cumulants of signals. Then estimate blindly the parameters of the three typical signals by Rife, DPT, Newton iteration and MAT algorithm. Finally, evaluate the reliability of the recognition and estimation results.The main work of the article is as follows.(1)For different modulation radar signals, this paper presents a new method to extract instantaneous frequency by instantaneous autocorrelation. According to the Intra-pulse features parameters such as correlation coefficient Rxy, normalized instantaneous frequency autocorrelation mean E and variance ? to select the same modulation type. Based on the improved higher-order cumulants algorithm, make the MPSK signals precise classified accurately.(2)Estimate the carrier frequency of conventional signal using the Rife algorithm and improve the low estimation precision nearby the quantitative frequency by optimized algorithm of frequency shift. The new joint algorithm presented based on DPT and Newton iteration improves the low estimation precision of DPT and large computation of two-dimensional search, which estimate the original frequency and frequency modulation coefficient of LFM signal.Blind estimate carrier frequency of the BPSK signal by the MAT algorithm and enhance the SNR to further improve the estimation accuracy of carrier frequency by a medium filter.(3)Evaluate the reliability of the recognition and estimation results. Establish a reference signal model and corre-cumulate with the observer signals.The linear regression fitting degree can verify the accuracy of recognition and the accuracy of parameter estimation.