Perfect Sampling on Blind Receivers Fading Channel with Unknown Channel State Information

This paper presents perfect sampling based on coupling from the past (CFTP) algorithms for blind detection about flat fading channel. The basic idea is to take CFTP as a protocol for finite-state Markov chain Monte Carlo (MCMC) method and by itself. The algorithm can determine the necessary runtime to convergence of state with unknown channel parameters by using the sandwiched CFTP scheme distribution, which allows for perfect (exact) and independent sampling from the desired stationary state distribution of the fading channels with unknown statistics so as to lower the computation complexity. On above, we generate samples from the likelihood function of the posterior distribution and extract relevant information. Simulation results are provided to demonstrate the excellent performance of the proposed blind receiver method.