Mimo Signal Detection Algorithms Based On Mcmc Methods

In recent years, as people increasingly require for high rate and high reliability of wireless communication, how to improve the transmission rate and reliability of the communication without increasing the bandwidth has become a technical difficulty of the modern communication systems. Multi-input Multi-output (MIMO) technique is such one which is able to increase the speed and reliability of the communication without increasing the bandwidth. It has always been the research focus of the modern communication techniques.Although the MIMO technique has the advantages, it also enhances the interference between the signals, thus greatly increases the dimensions of the signals. So the design of the detection algorithms has been more difficult at the receiver. How to design algorithms which have low computational complexity and near optimum performance has become a hot research topic of MIMO signal detection. Howere, it’s almost impossible to simultaneously obtain low computational complexity and near optimal performance for most of the traditional MIMO signal detection algorithms. Markov chain Monte Carlo (MCMC) methods, howere, not only can approach the optimal performance, but also its computational complexity is not significantly increasing as the number of the antennas and the size of the signal space. Actually, it is linearly increasing with the number of the iterations.The thesis analyzes the MIMO signal detection algorithms based on MCMC methods in three different channel conditions. Firstly, the channel status is completely known and the channel is the frequency flat faded, and the MCMC MIMO signal detection algorithms are relatively simple, which obtain samples from conditional probability function, and the type of the sampling method determines the final performance. Secondly, the channel status is not completely known and the channel is frequency flat faded, and the MCMC MIMO signal detection algorithms can obtain the initial estimation of the channel by utilizing pilot sequences, then the algorithms calculate likelihood ratio according to the estimated channel state and the estimated error. Finally, we implement turbo receiver between the MCMC detector and channel decoder. Thirdly, the channel status is completely unknown and the channel is frequency selective faded, and the MCMC MIMO signal detection algorithms transferms the Multi-input Single-output (MISO) system into Single-input Multi-output (SIMO) system by utilizing the characteristics of the Orthogonal Space-time Block Coding (OSTBC) and then performs blind channel estimation and symbol detection. Here we also implement the turbo receiver between the detector and the channel decoder.Previous works mentionded that the MCMC method could encounter the problem of the performance degradation under high signal-to-noise ratio. This thesis shows that we can overcome this problem by adding the redundancy check bits into the packets and stopping the iterative process when the correct detection probability of packets reaches a certain level. This method does not greatly increase the complexity compared with the method which blind increase the number of iterations.