K. Yen

Antenna-diversity-assisted genetic-algorithm-based multiuser detection schemes for synchronous CDMA systems

A spatial diversity reception assisted multiuser code-division multiple-access detector based on genetic algorithms (GAs) is proposed. Two different GA-based individual-selection strategies are considered. In our first approach, the so-called individuals of the GA are selected for further exploitation, based purely on the sum of their corresponding figures of merit evaluated for the individual antennas. According to our second strategy, the GAs individuals are selected based on the concept of the so-called Pareto optimality, which uses the information from the individual antennas independently. Computer simulations showed that the GAs employing the latter strategy achieve a lower bit-error rate as compared to the former strategy. For a 15-user GA-assisted system employing a spreading factor of 31, a complexity reduction factor of 81 was achieved at a performance identical to that of the optimum multiuser detector using a full search.

M-ary coded modulation assisted genetic algorithm based multiuser detection for CDMA systems

In this contribution we propose a novel M-ary coded modulation assisted genetic algorithm based multiuser detection (CM-GA-MUD) scheme for synchronous CDMA systems. The performance of the proposed scheme was investigated using quadrature-phase-shift-keying (QPSK), 8-level PSK and 16-level quadrature amplitude modulation (16QAM) when communicating over AWGN and narrowband Rayleigh fading channels. When compared with the optimum MUD scheme, the GA-MUD subsystem is capable or reducing the computational complexity significantly. On the other hand, the CM subsystem is capable of obtaining considerable coding gains despite being fed with sub-optimal information provided by the GA-MUD output.

Genetic algorithm based antenna diversity assisted multiuser detection for synchronous CDMA systems

A spatial diversity reception assisted multiuser CDMA detector based on genetic algorithms (GA) is proposed. Two different diversity selection strategies are considered. In our first approach the so-called individuals of the GA are chosen for selection based purely on the sum of their corresponding figures of merit associated with the individual antennas, an approach which is analogous to employing the conventional log-likelihood function (LLF) for diversity reception. According to our second strategy, the individuals are chosen for selection based on the concept of the so-called Pareto optimality, which uses the information from the individual antennas independently. Computer Simulations showed that the GA employing the latter. strategy achieve a lower BER as compared to the former approach.