Efficient complementary sequences-based architectures and their application to ranging measurements

Resum

In the last decades, ranging systems have benefited from advances in the wireless communication field, as multiple access techniques or near-far mitigation algorithms. In CDMAbased (Code-Division Multiple-Access) ranging systems, the properties of the spreading sequence used play a key role on the development of high-precision ranging measurements. This thesis proposes novel efficient generation/correlation architectures of Complementary Sets of Sequences (CSS) and sequences derived from them, as Loosely Synchronized (LS) and Generalized Pairwise Complementary (GPC) sequences. We consider the term efficient applicable whether the proposed architectures requires less operations per input sample in comparison with a straighforward implementation (a Tapped-Delay Line implementation). The contributions of the thesis can be divided into two stages: Firstly, we generalize the efficient generation/correlation architectures for binary CSS, derived in previous works, to the multilevel (real-valued) alphabet by using multilevel Hadamard matrices. This approach has two advantages: on the one hand the increase of the feasible lengths that the architecture is capable to generate/correlate, and the elimination of the previous limitations in the number of sequences of the set; on the other hand, under certain conditions, the generalized architectures allow to particularize their inner structure to efficiently generate/correlate binary CSS with more different lengths than the ones feasible with previous efficient algorithms. Secondly, based on the proposed architectures, we provide novel algorithms for LS and GPC sequences that reduce the number of operations per input sample. Finally, we do a comparative analysis of the performance of LS and Kasami sequences in a novel UWB indoor positioning system