Estimation and tracking of time-varying channels in OFDM systems

We propose a method for channel estimation in orthogonal frequency division multiplexing (OFDM) systems. Unlike the conventional sample-spaced and sub-sample-spaced methods, which target the taps of an equivalent discrete-delay channel response, this method targets the physical propagation paths in a continuous-delay domain. Numerical results quantify its benefits over the conventional least-squares, orthogonal matching pursuit, and mixed-norm methods. Assuming that channel is slowly varying from one OFDM block to another, we also develop an adaptive algorithm for channel tracking. Performance is illustrated through simulation.