Study On Multi-User Detection In MIMO System With Quantum Genetic Sphere Algorithm

Multi-user detection (MUD), which decides to decoder perform of MIMO communication system, is one of the important technologies in modern high-speed wireless digital communication, and its detection efficiency directly affects the system’s data throughput. How to improve the multi-user detection technology to improve detecting rate, lower error rate and construct high speed and high reliability multi-user detector is one problem to be solved in the field of wireless digital communication. Therefore, studying multi-user detection algorithm to improve its efficiency will be a thing with great important practical significance.Principle of MIMO systems has been summarized, followed by in-depth discussion of the mathematical model of MIMO communication systems with the example of 16-QAM modulation CDMA communication system. Then the MUD of MIMO systems has been constructed by adopting the traditional sphere decoding (SD). Simulation results indicate that the randomness of transmitted signal, environmental noise and various kinds of interference noise make the receiver of MIMO system detecting the original signals from transmitters more accurately at the same time difficulty.So a binary coded genetic algorithm (GA) for MUD of MIMO system is proposed in the paper. The new binary coded GA multiuser detector maps the constellation points into binary strings with fixed length by adopting the discrete characteristic of signal constellations. Genetic algorithm realizes its strong ability of global parallel search for complex multi-variable optimization problems due to its parallel processing, adaptive and nonlinear mapping, thus resulting in high accuracy in decoding. Simulation results indicate that the error rate performance obtained from GA detector has considerably improved compared to zero forcing (ZF) detector method. However, due to the search process of GA depends on the algorithm cycle steps and the size of initial population, the single decoding time will increase with the number of user increases exponentially.A novel sphere decoding algorithm is developed in the paper combining with quantum algorithm and GA in order to overcome the shortcomings of the proposed binary coded GA detector. The quantum genetic sphere decoding algorithm (QGASD) combines the parallelism of quantum computing and genetic algorithms overall, so that the search of each layer do not need to try several times repeatedly. And because the sphere algorithm can be summarized as a complex multi-dimensional search for a simple one-dimensional search, and determine the solution interval of each dimension, the new QGASD can avoid a large number of complex matrix operations in searching process. So its search efficiency may be greatly improved. Simulation results show that the performances of QGASD such as the decoding speed, error rate and computational complexity are greatly improved in comparison with the techniques of the original SD, GA detector and ZF detector.