Achieving Code Synchronization In Direct-Sequence Spread Spectrum System Applications Using Recurrent Neural Networks

Code synchronization is the most important matter of concern in all applications involving direct sequence spread spectrum system, such as CDMA. Conventional synchronization techniques are primarily based on the auto/cross correlation properties of the spreading codes. The objective of the present paper is to propose a new method for fast and reliable code synchronization based on the abilities of recurrent neural networks in pattern association problems. One major benefit of this method is the independence of its performance on correlation characteristics of the spreading codes. The feasibility of the proposed scheme in achieving synchronization between transmitter and receiver in the presence of noise and interference are considered and the results are presented