Iterative message-passing-based algorithms to detect spreading codes

This thesis tackles the issue of the rapid acquisition of spreading codes in Direct-Sequence Spread-Spectrum (DS/SS) communication systems. In particular, a new algorithm is proposed that exploits the experience of the iterative decoding of modern codes (LDPC and turbo codes) to detect these sequences. This new method is a Message-Passing-based algorithm. Specifically, instead of correlating the received signal with local replicas of the transmitted linear feedback shift register (LFSR) sequence, an iterative Message-Passing algorithm is implemented to be run on a loopy graph. In particular, these graphical models are designed by manipulating the generating polynomial structure of the considered LFSR sequence. Therefore, this contribution is a detailed analysis of the detection technique based on Message-Passing algorithms to acquire m-Sequences and Gold codes. More in detail, a unified treatment to design and implement a specific set of graphical models for these codes is reported. A theoretical study on the acquisition time performance and their comparison to the standard algorithms (full-parallel, simple-serial, and hybrid searches) is done. A preliminary architectural design is also provided. Finally, the analysis is also enriched by comparing this new technique to the standard algorithms in terms of computational complexity and (missed/wrong/correct) acquisition probabilities as derived by simulations.