In this article, we propose a process to minimize the Multi-User Interference (MUI) caused by Multiple Access (MA) in a Time Hopping Ultra Wide Band (TH-UWB) system with No Line-Of-Sight (NLOS) multipath channel in a residential environment IEEE 802.15.4a Channel Model 2 (CM2). We discuss both synchronous and asynchronous channels with an Additive White Gaussian Noise (AWGN). The choice of codes allowing multiple access is crucial for minimizing the MUI and to improve the system performance. Recently much works has been done in order to optimize performance by code selection [Bravo et al., 1999; Moreno & Marić, 2000; Guvenc & Arslan, 2004]. This paper is a contribution to this research axis. The minimization of the MUI is formulated as a Combinatorial Optimization Problem (COP) and solved by means of an efficient generic optimization method called Guided Local Search (GLS) [Voudouris, 1997]. In this context, we exploit the richness of chaotic sequences with respect to topologic properties. Simulation results based on the Binary Error Rate (BER) show that optimized chaotic sequences outperforms the classical code sequences.
Parallelizing Combinatorial Optimization Heuristics with GPUs