Markov Chain Monte Carlo Algorithms And Its Application In MIMO Detection

With the rapid development of multimedia services and the ever increasing demand for high-speed data transmission, the efficient techniques which improve the scare wireless spectrum resource exploitation play an important role in wireless communication system. Due to its advantages in enhancing the spectrum efficiency without increasing the system bandwidth, multiple input and multiple output (MIMO) technique, more specifically, the effective detection algorithm with respect to the space-time coding scheme, have become one of the key enabling technique to support broadband communication. Therefore, MIMO and the related detection algorithm have always been the research focus in recent years.Recently, it is proposed to apply the Markov Chain Monte Carlo (MCMC) method to solve the MIMO detection problem with low calculation complexity. After a brief introduction of the involved fundamental theory and essence in MCMC method, the MCMC based estimation over AWGN channel and the MCMC based MIMO detection, as well as the achieved detection performance, are presented as illustrative examples to illuminate the basic idea of the MCMC-based detection scheme, and to argue for its feasiblity. It is unveiled through numerical simulation results that, the MCMC-based detection algorithm may achieve reasonable reliability, which is superior than that realized by using the recursive MMSE-DFE detection algorithm, while the calculation complexity only depends on the number of iterative detection.It is shown that the perfect space-time block code (PSTBC) outperforms other space-time coding scheme, albeit its detection algorithms need further investigation. In this thesis, by transforming the PSTBC code into the equivalent VBLAST model, the MCMC based PSTBC detection algorithm is addressed. It is validated through simulations that, the performance of MCMC algorithm outperform the performance of Fano detection algorithms. In addition, there will be a limited increase in computational complexity with the increase in modulation alphabet size.Moreover, the MCMC-based joint iterative detection and decoding technique for Space-Time code and error control coding scheme is addressed in this thesis. Our analysis shows that, the MCMC-based joint iterative detection and decoding algorithm exihibits the diverging problem within high SNR region. And the diverging problem could be explained by the fact that, within high SNR region, some of the transition probabilities in the underlying Markov chain become problematic, as a result, the MCMC tends to be trapped in some state. In order to overcome this problem, it is proposed to increase the number of samples to improve the convergence performance. And the simulation results are presented to validate that increasing the number of samples is able to improve the joint iterative detection and decoding within high SNR region, and the only paid cost is some increase in the calculation complexity.