Researches On Key Technology Of Blind Despreading For DS-CDMA Signals

Direct sequence code division multiple access(DS-CDMA) signals have found wide applications in military and civil communications, which makes the non-cooperative reception of DS-CDMA signals of great significance. Taking the non-cooperative communication as the application background, this thesis mainly researches on the blind despreading of DS-CDMA signals. The main contributions are as follows:1.An improved method is proposed to estimate the user's number of DS-CDMA system, using the number of the signal's eigenvalue in autocorrelation matrix. An adaptive threshold is introduced for discriminating signal eigenvalues from the noise ones in the method. Simulation results show that the correct estimation rate of user number is larger than 99%, when SNR is -7dB.2.A synchronization method based on eigenvalues variation in terms of the observation window's shifts is given. Simulation demonstrates that the symbol synchronization position can be estimated simultaneously for all users, and the complexity are remained relatively constant with the increase of the users.3. The present algorithms for blind synchronization and blind spreading sequence estimation of short-code DS-CDMA signals have high computational complexity. Aiming at this problem, a novel low complexity algorithm is proposed. Based on the structure of the signal correlation matrix, the algorithm firstly sets proper length for observation windows, then estimates the timing offsets using spectral norm and obtains the spreading codes by eigenvalue decomposition in an iterative manner. Simulation shows that multi-users' timing offsets and spreading codes, in both synchronous and asynchronous scenario, can be estimated with a high correction rate and good real-time performance by this algorithm.