Receiver design for MIMO OFDM transmission over time variant channels

This paper considers receiver design for space time block coded MIMO OFDM transmission over frequency and time selective channels. The receiver employs the expectation-maximization algorithm for joint channel and data recovery. It makes collective use of the data and channel constraints of the communication problem. The data constraints include pilots, the cyclic prefix, the finite alphabet constraint, and space-time block coding. The channel constraints include the finite delay spread and frequency and time correlation as well as spatial correlation. The channel estimation part of the receiver boils down to an EM-based forward-backward Kalman filter. To avoid the latency and storage associated with smoothing, we introduce a forward-only Kalman that performs channel (and data) recovery with no latency. Simulations show that the receiver outperforms other least-squares based iterative receivers.