An EM based frequency domain channel estimation algorithm for multi-access OFDM systems

Channel estimation is an important prerequisite for receiver design. In this paper we present a semi-blind low complexity frequency domain based channel estimation algorithm for multi-access Orthogonal Frequency Division Multiplexing (OFDM) systems. Our algorithm is based on eigenvalues interpolation and makes a collective use of data and channel constraints. We exploit these constraints to derive a frequency domain maximum a posteriori (MAP) channel estimator. Furthermore, we develop a data aided (expectation maximization based) estimator incorporating frequency correlation information. The estimator is further enhanced by utilizing the time correlation information through a forward backward (FB) Kalman filter. We also explore various implementation for the FB Kalman filter. The simulation results are provided validating the applicability of the proposed algorithm.