PN Sequence Period Estimation Of DS/SS Signals Based On The Improved Cepstrum Approach

With the development of electronic and information war, blind de-spread of spread spectrum signal under the condition of noncooperative communication becomes the research focus. Meanwhile, blind inspection and blind estimation of pseudo-random code period in spread spectrum signal is the precondition for the realization of blind de-spread.This treatise respectively utilizing conventional spectrum estimation and modern spectrum estimation to improve traditional cepstrum approach. Through theoretical analysis and simulation, an effective method of the estimation of pseudo-random code period is obtained in the condition of low SNR(Signal Noise Rate).Firstly, this treatise summarizes and analyzes spread spectrum and DS system, studies methodology- power spectrum estimation. The method of period chart and the three improved methods in conventional power spectrum estimation are analyzed in detail. The theory of AR(Auto Regression) parameter model in modern power spectrum estimation, including the way of coefficient calculation and decision of order, are discussed.Secondly, conventional power spectrum estimation is applied to cepstrum approach and traditional cepstrum approach is improved by three methods. The first one is to employ cross-correlation function to estimate power spectrum. The second one is to use windowing autocorrelation function. The third one is to revise spectrum of cepstrum approach. By the simulation, it is proved that improved cepstrum approach is better than traditional one, because its peak value is more extrusive and the line-spectrum is more smooth and it is beneficial to search the peak value. As to the signal modulated by short code, SNR can be -20.1dB and it is superior to -18dB in existing methods. Moreover, the period of long code signal can also be estimated by improved cepstrum approach based on delay multiplication. By this method, SNR attains -13.1dB so that it solves the problem that cepstrum approach cannot estimate the signal modulated by long code in current literature. In addition, by emulation mode, it is obtained that the shorter the code, the better the effect of cepstrum approach estimation and the more the data, the lower the signal-to-noise ratio tolerance.Finally, modern power spectrum estimation is applied to cepstrum approach. Utilize the arithmetics of Yule-Walker and Burg to estimate power spectrum, calculate cepstrum and then revise its spectrum. By the simulation, it is manifested that cepstrum approach based on modern spectrum estimation is better than traditional spectrum estimation. Meanwhile, this paper concludes empirical rules to decide the order in AR model of cepstrum approach.