The Non-periodic Dsss Signal Detection And Parameter Estimation

With the increase of the demand of communications security, anti-jamming capabilities, bandwidth utilization and signal secrecy, an effective signal form——non-periodic Direct Sequence Spread Spectrum (DSSS) signal has been widely used in the civilian and military communications. One cycle of the spread-spectrum sequence includes some cycles of the information symbol. That is, different information symbol codes correspond to different spread-spectrum PN codes. So the non-periodic DSSS signal is different from the general DS signal. Since different information codes are independent of each other and the correlation of the PN code is undermined, the detection and the parameter estimation of non-periodic DSSS signal become a hotspot in the non-cooperative communication.The dissertation focuses on the research of the detection and the parameter estimation for the non-periodic DSSS signal based on the theory of high-order statistics and cycle-related spectral.First of all, the non-periodic DSSS communication system, signal model, high-order statistics and cycle-related spectral theory are introduced.Then, through analyzing the non-periodic DSSS signal's second-order moment ( m2_s(τ) ), fourth-order moment chips( m4_s(0,0,0), m4_s(0,0,τ) , m4_s(0,τ,τ) , m4_s(τ,τ,τ) ) and the received signal's second-order moment ( m2_x(τ) ), fourth-order moment chips( m4_x(0,0,0),m 4x(0 ,0,τ),m4_x(0,τ,τ),m4_x(τ) ,τ,τ), two new detection methods(EJJ, ECSJS) based on the second-order moment and the quadratic fourth-order moment chip are proposed, respectively. According to the fact that the received signal's fourth-order moment chips have the information of carrier frequency, the detection method(ECSJP) in frequency based on quadratic fourth-order moment chip is proposed. Because the cycle-related spectrum ( S_y~α(f)) of the fourth-order moment chip ( m 4x(0 ,0,τ)) has the information of carrier frequency, the new detection method(SJXH) based on the cycle-related spectrum is proposed. Experimental results show that the SJXH method has the best performance on non-periodic DSSS signal detection. When the detection probability is 90%, the SJXH method has better performance than the ECSJS method to improve SNR about 0.5dB, compared with ECSJP method to improve the SNR about 1.5dB, compared with EJJ method to improve the SNR about 3dB, compared with spectrum-reprocessing method to improve the SNR about 10dB and compared with double frequency method to improve the SNR about 15dB.After all, according to the fact that the fourth-order moment chip and the cycle-related spectral have the information of carrier frequency, two kinds of carrier frequency estimation methods(ECSJSP,SJXHP) of non-periodic DSSS signal based on quadratic fourth-order moment chip and the fourth-order moment chip's cycle-related spectral are proposed. Through analyzing the received signal's fourth-order moment reprocessing-spectrum, an estimation method for PN sequence period and for PN chip rate are proposed, respectively. Experimental results show that when the normalized estimation relative error(NERE) of carrier frequency estimation is smaller than 10?2 , the ECSJSP method and SJXHP method have better performance than the double frequency method to improve SNR about 12dB and 13dB, respectively. But cepstrum method can not estimate the non-periodic DSSS signal's carrier frequency. When the normalized estimation relative error(NERE) of PN sequence period estimation is smaller than 10?3 , the proposed estimation method for PN sequence period has better performance than the time-domain correlation method to improve SNR about 4dB, compared with spectrum-reprocessing method to improve the SNR about 5dB. When the NERE of PN chip rate estimation is smaller than 10?2 , the proposed estimation method for PN sequence period has better performance than the time-domain correlation method and spectrum-reprocessing method to improve SNR about 2dB, compared with cepstrum method to improve the SNR more than 6dB.