Bayesian Monte Carlo signal processing for wireless communication

Optimal multi-parameter estimation for wireless communication is addressed in this work. The Bayesian Monte Carlo signal processing as a novel approach for Bayesian computation is introduced. Two families of Monte Carlo signal processing methodologies are outlined, namely, Markov chain Monte Carlo (MCMC) methods and sequential Monte Carlo (SMC) methods. The main contribution of this work is to develop Bayesian Monte Carlo solutions to two classes of problems: optimal receiver design for wireless transmission and mobility management in cellular networks.; Within the topic of optimal receiver design, specific problems addressed here include sequence detection for GMSK signal over unknown multipath channel; multiuser detection for asynchronous CDMA system in the presence of unknown multipath channel, narrowband interference and other out-cell interference; multiuser detection for multi-carrier CDMA system with transmission diversity; sequence estimation for OFDM system in frequency-selective fading channels. Being soft-input and soft-output in nature, Monte Carlo technique is well suited for iterative (turbo) processing for a channel coded system. Bayesian detectors as well as Bayesian turbo receivers are therefore derived for the above systems without explicit channel estimation.; Mobility tracking and handoff detection as two important components are addressed for mobility management in cellular networks. The system is first modeled as non-linear dynamic system, the sequential Monte Carlo (SMC) methods are then employed to compute the on-line estimation and detection. In addition, novel handoff schemes are derived based on different optimization criteria.; It is demonstrated through simulation that the proposed Bayesian Monte Carlo techniques can achieve a close-to-optimal performance at a reasonable computational cost.