Adaptive Hybrid Genetic Algorithm Parallel Interference Cancellation High Rate Multi-User Detection for Dual Rate W-CDMA Mobile Communications.
Liangfang Ni,Sidan Du,Baoyu Zheng
DOI: https://doi.org/10.1093/ietcom/e90-b.7.1692
2007-01-01
IEICE Transactions on Communications
Abstract:Adaptive hybrid genetic algorithm concatenated with improved parallel interference cancellation, i.e. adaptive hybrid genetic algorithm parallel interference cancellation (AHGAPIC) was proposed. A study is conducted on the application of AHGAPIC to soft decoding high rate multi-user detection with diversity reception for dual-rate wideband DS-CDMA spread spectrum communications, aiming to mitigate the effect of multiple access interference. The relevant research has revealed that the local search capability of hybrid genetic algorithm (HGA) is still not good enough. Therefore, first, two evolutionary operations, i.e. inversion and insertion are merged into HGA to constitute a novel algorithm. With its moderate local search capability, this new algorithm can search for the global optimum region according to the information entropy, and then it is made adaptively vary its probabilities of crossover and mutation depending on the fitness values of the solutions to form the adaptive hybrid genetic algorithm (AHGA). Second, AHGA is utilized to effectively identify the better and better binary string to maximize the log-likelihood function of dual-rate multi-user detection. As AHGA converges to the optimum region, the control factor of the improved parallel interference cancellation (IPIC) detector is set to be the ratio of the average fitness value to the maximum fitness value of the population of AHGA. Finally, equipped with both the control factor and the binary string with the maximum fitness value as the initial data, the IPIC detector can rapidly find out the approximately optimum soft decoding vector. Then, it can obtain the approximately global optimum estimate point on the basis of the soft decoding rule, corresponding to the transmitted data bits. A lower bound of computational complexity has been achieved through simulations and qualitative analyses. The property of the proposed algorithm to converge rapidly leads to lower computational complexity. Emulation results have shown that the AHGAPIC soft decoding high rate multi-user detector is superior to other suboptimum detectors considered in this paper in terms of two points. They are the mitigation of multiple access interference and the resistance to near-far effects. Its performance is close to the sequential group optimum multi-user detector but with a shorter time delay.