Data Detection and Fixed Symbol-Timing Estimation in Flat Fading Channels

This paper deals with data detection in flat fading channels where the symbol sampling instants are unknown. Optimal solution for the problem of estimation of synchronization parameters, such as symbol-timing, in communication systems is analytically intractable, hence most of the approaches found in the literature are based on approximate and heuristic methods. We approach the problem from Bayesian perspective, and jointly estimate the symbol timing and the transmitted symbols using particle filtering techniques. In addressing the problem, we consider the symbol-timing to be fixed and recursively approximate its posterior probability density by a beta distribution function. Moreover, the channel fading is taken as a Rayleigh distributed process and is modeled by a first order autoregressive process. Such modeling yields to the representation of the problem by a dynamic system in a state-space form that allows the application of particle filtering methods for obtaining a near optimal estimation of the transmitted symbols, the symbol timing and the complex amplitude of the fading channel. Computer simulations demonstrate that the proposed algorithm shows excellent performance.