Robust Multiuser Detection Based on Particle Swarm Optimization for CDMA Communications

• Main Authors: Jhao-Li Meng, 孟昭立
• Other Authors: Ann-Chen Chang
• Format: Others
• Language: zh-TW
• Published: 2011
• Online Access: http://ndltd.ncl.edu.tw/handle/35773233193577745900

碩士 === 嶺東科技大學 === 資訊科技應用研究所 === 99 === Spread spectrum-based code division multiple access (CDMA), has taken on a significant role in cellular and personal communications because of such characteristics as potential capacity increases over anti-multipath capabilities, soft capacity, and soft handoff. Multiple-access allows multiple users to share limited resources such as frequency (bandwidth) and time. Direct-sequence CDMA (DS-CDMA) is the most popular of CDMA techniques. However, there are two key limits to present DS-CDMA systems: multiple-access interference (MAI) and near-far effect. In DS-CDMA systems, the multiuser detection (MUD) has been used to combat MAI and near-far effect. The maximum-likelihood (ML) detection can effectively combat the MAI problem; however, its complexity increases exponentially with the number of users. Linear multiuser detection has lower computational complexity, but it cannot provide sufficient performance. In this thesis, we investigate the MUD strategy using particle swarm optimization (PSO) algorithm, which has lower computation complexity, to search the maximum likelihood results. Recently, PSO algorithm is a population based heuristic global optimization technology has been proposed. In PSO algorithm, the individual is called particle which using identical inertia weight in velocity updating and has a superior convergence speed compared with other evolutionary algorithm. However, it cannot be guaranteed that all the particles converge to a global best position due to high convergence speed. Therefore, this thesis presents a modified PSO algorithm with velocity updating by using Newton method to improve the accuracy of optimal solution searching and convergence speed for MUD. Antenna array has been widely implemented in practical mobile communication system because of its directional radiation pattern can improve the performance of signal receiving. Due to the estimation of direction of arrival (DOA) of signal has difficulties in nature, adaptive array beamforming always exists the problem of pointing error. To mitigate the performance degradation due to pointing error and finite sample effects, this thesis also deals with eigenspace-based (ESB) beamforming based on the decision-directed (DD) correction with robust capability. However, it is shown that the output SINR of the ESB beamformer is degraded when the pointing error is relatively larger. As a matter of fact, the received signal obtained from the reference element of the uniform linear array is only time-domain data. In conjugation with DD and PSO-based chip estimation procedure, the proposed approach can be used to form a robust DD-ESB beamformer without any specific training sequence. Simulation results show that the proposed DD-ESB beamformer provides more robust capabilities than the existing ESB, RESB, and DD-ESB beamformers, respectively.