Discrimination Analysis And Parameter Estimation Of LPI Radar Signals

Electronic surveillance system for radar signal interception process can be broadlydescribed as: detecting radar signals; analyzing of the radar signals; identifying radarsignals; locating enemy radar according to the results of the previous analysis, andanalysis of the local combat sequences, as our front-line intelligence. The main work forLPI radar intercepts performance verification specifically including the followingworks:1、Firstly, make modeling and analysis on several classic radar signals, thenintroduce curve fitting and fractional Fourier transform which will be used in this study.After that, we deeply analysis phase unwrapping, which remedies the shortage oftraditional phase unwrapping algorithm, that requires a priori knowledge of the signal.In signal detection and identification process, we firstly estimate center frequency andbandwidth of the signal roughly, then we do frequency spectrum shifting according tothe sampling frequency, finally unwrap the phase to get the true phase.2、Introduce two algorithms of time-frequency analysis, the short time Fouriertransform and Wigner-Ville distribution, and do time-frequency analysis on LPI radarsignal. After that, we use smooth and pseudo Wigner-Ville distribution (SPWD) andpseudo Wigner-Ville distribution (PWD) which are well improved in suppressing thecross-term to analyze time-frequency of FSK.3、According to the introduced phase unwrapping algorithm, the instantaneousphase of the radar signal is found, then process N-order differential on the signal.According to the characteristic of the time-frequency curves, we can recognize thesignals by using tools as curve fitting, peak detection, and so on. Lastly we can examthe feasibility of the program by Monte Carlo simulation.4、Firstly estimate parameter of LFM signal. First of all, dimensionless normalizetime domain and frequency domain of the signal. Secondly, do FRFT on the signal toget the spectrum distribution figure. Thirdly, search the peak of the spectrum to estimatethe parameter of the signal, and exam the feasibility through simulation. Then we doparameter estimation on2FSK signal. Firstly, we introduce one method of parameter estimation, which is based on the curve one-order differenced from time-frequencycurve; peak detection is used in this method. Due to the presence of noise, the error isslightly large. After that, the paper studies a method by constructing a new signal toestimate the parameter. This new method narrowed the estimation range using theresults of the former method, and then searches the trip points of the signal by findingthe maximum and minimum. Through this method, the error is smaller, and the methodis helpful in engineering steering.This paper do much simulation after each part of introduction and studying, theresults show that the proposed method is feasible and effective, and the methods havegood performance in the premise of a certain signal-noise ratio.