Study On Parameter Estimation Of The Non-Cooperated DS Signal Based On Spectral Correlation

The problem of estimating DS signals has been of great research interest with the development of the field of wind-band weak signal processing and the field of communication antagonism because of the DS signals'strong anti-interference ability, low probability of being intercepted and outstanding multi-access communication ability. In 1980s, people start to study them, through these years, many detection and estimation methods has been developed. They are almost based on the theory of conventional detection and estimation. So when the signal to noise ratios of the received signals becomes lower, the performance of these methods take a turn for the worse, and these methods don't work.The study of DS signals in this article broke through the conventional detection and estimation methods and modeled the DS signals as a cyclostationary random process. In this case, it can be analyzed by the theory of spectral correlation, and then we can finish the parameter estimation of DS signals by its characteristic in the field of spectral correlation.DS signal is twice modulated with the rapid pseudorandom noise sequence and the carrier. So it appears cyclostationary. In the field of spectral correlation, DS signals appears its characteristic as fallows: for the amplitude of spectral correlation, the information of carrier frequency and PN code rate are contained on the spectrum frequency equal to zero sheet, and also the PN code rate is contained on the sheet of spectrum frequency equal to carrier frequency; For the phase of spectral correlation, the information of carrier initial phase is contained on the sheet of cyclic frequency equal to double carrier frequency. Therefore, the three parameters of DS signals mentioned above can be estimated by these characters.The spectral correlation density must be estimated before the DS signal parameter estimation. In order to reduce the effect of random noise, we provide a improved method to estimate the spectral correlation density which is based on the average algorithm. It can make the spectral line be more obvious, and improved the SNR limit when estimating these parameters.