Research On Detection And Parameter Estimation Technology For Direct Sequence Spread Spectrum Signals At Low SNR

Direct sequence spread spectrum(DSSS)communication,as a main working mode of spread spectrum communication,is widely used in military and civil communication due to its good anti-interference and concealment.In the field of electronic reconnaissance and communication countermeasures,how to realize the detection and parameter estimation of DSSS signals under low signal to noise ratio(SNR)is always an important topic for scholars at home and abroad.In this paper,DSSS/BPSK is used as the main research signal,and some methods of DSSS signal detection and parameter estimation are studied.In the aspect of detection,the auto-correlation algorithm and the fourth-order cumulant slice algorithm are researched.Based on the traditional algorithm,wavelet decomposition and delay multiplication techniques are used to improve the two algorithms.Two improved algorithms of auto-correlation and fourth-order cumulant based on wavelet decomposition and delay multiplication are proposed.The simulation shows that the improved algorithms in the integer multiples of pseudo-noise(PN)code period can appear obvious spectral peaks which make it easier to detect the presence or absence of a signal.When the SNR is no less than-12 dB,the average correct detection probability of DSSS signal can reach more than 90%,which effectively improves the detection performance of DSSS signal and achieves the accurate estimation of PN code under low SNR.In terms of parameter estimation,this paper deeply studies the delay multiplication algorithm and the cyclic spectrum algorithm,analyzes its theory,and simulates the performance of the algorithm.The delay multiplication algorithm is mainly used to estimate the PN code rate,while the cyclic spectrum algorithm can simultaneously estimate the carrier frequency and PN code rate of DSSS signals.In addition,the autocorrelation method and the fourth-order cumulant slice method for DSSS signal detection can also obtain the information of the PN code period while detecting the signal,and the fourth-order cumulant slice method can also estimate the carrier frequency of the signal.In this paper,the performance of these four parameter estimation methods is simulated and compared,and the merits and drawbacks of each algorithm are analyzed combined with the simulation results.In PN code sequence estimation,this paper studies the eigen value decomposition(EVD)and variable step-size neural network algorithm for single user,and Independent Component Algorithm(ICA)for multi-user.On the basis of analyzing the traditional algorithm,aiming at the phase ambiguity problem of EVD in asynchronous situation,the double window method is used to improve it,and the improved algorithm is simulated and analyzed.In order to solve the problem of large amount of computation and high complexity of EVD algorithm,variable step-size neural network algorithm is introduced as a fast calculation method of EVD,and the performance of two PN code sequence estimation algorithms is compared and analyzed.Finally,this paper also uses ICA technology to study the multi-user PN code sequence estimation.On the basis of fully considering the application background,Kernel Entropy Component Analysis(KECA)is introduced to replace Principle Component Analysis(PCA)for signal whitening to optimize FastICA algorithm.A blind estimation algorithm based on KECA-ICA is proposed,which makes the final estimation result more accurate and reliable.