Joint channel estimation and Markov Chain Monte Carlo detection for frequency-selective channels

In this paper, we develop a novel approach for joint channel estimation and Markov Chain Monte Carlo (MCMC) detection for time-varying frequency-selective channels. First, we propose a sequential channel estimation (SCE) MCMC algorithm that combines an MCMC algorithm for data detection, and an adaptive least mean square (LMS) algorithm for channel tracking, in a sequential fashion. Then we develop a stochastic expectation maximization (SEM) MCMC algorithm that takes advantage of both the MCMC approach and the EM algorithm to find jointly important samples of the transmitted data and channel impulse response (CIR). The proposed algorithms provide a low-complexity means to approximate the optimal maximum a posterior (MAP) detection in a statistical fashion and are applicable to channels with long memory. Excellent behavior of the proposed algorithms is presented using both synthetic channels and real data collected from actual underwater acoustic experiments.