Research On Multi-user Detection Technologies In CDMA Communication System

In the CDMA communication system, different users are distinguished by Spread-Spectrum Codes. It's very difficult to keep completely orthogonal between Spread-Spectrum Codes, which leads to interference between different users in the same channel;this is called Multiple Access Interference (MAI). When the distance from users to the base station is different, Near-Far Problem will come. Traditionally, Match filters are used to extract the information of each user according to Direct Sequence Spread Spectrum theory, but cannot solve the problem of MAI and Near-Far Problem.Multi-User Detection technologies is based on the traditional detection technologies, it utilizes the information of all users to extract the object user from other users, which can decrease MAI and has better Near-Far Resistance. So the resource of frequency spectrum is fully utilized, and the capacity of the system is enlarged. The optimum multi-user detector proposed by Verdu is composed of match filters and Maximum Likelihood Sequential Estimation detector that is based on the Viterbi algorithm and has the lowest Bit-Error-Rate (BER) and best Near-Far Resistance. But the computational complexity is increased exponentially when the number of the users is increased, which becomes NP-Hard.In recent years, a new intelligent algorithm called Particle Swarm Optimization (PSO) is used to solve the NP-Hard, and it is testified that PSO has more simple computational complexity and faster convergence than genetic algorithm (GA), but has the same or better effect. In this paper, two kinds of sub-optimum detector basedon the standard binary particle swarm optimization is proposed------the multi-userdetector based on particle swarm optimization with maintenance of diversity;The multi-user detector based on clone selection particle swarm optimization. Computer simulation results have shown that the two kinds of detectors are superior to the detector based on the standard binary particle swarm optimization in BER and convergence. In addition, a multi-user detector based on new parallel particle swarm optimization is proposed in this paper, computer simulation results have shown that the detector based on new parallel particle swarm optimization is superior to the detector based on parallel particle swarm optimization in BER and convergence.