A Kalman Filtering Approach to Joint MIMO Channel Tracking and Symbol Decoding for Orthogonal Space-Time Block Codes

In this paper, we study the problem of MIMO channel tracking for communication systems that use orthogonal space-time block codes as the underlying space-time coding scheme. For such systems, we develop a two-step channel tracking algorithm. As the first step, our proposed method uses Kalman filtering to obtain an initial channel estimate, for each block, based on the channel estimate obtained for the previous block. In the second step, we use an iterative decision-directed method to refine the initial channel estimate obtained in the first step. We show that due to specific structure of OSTBC, both steps can be significantly simplified. Simulation results show that, for a training repetition period of 10 blocks, this algorithm can have a performance within 1 dB, in terms of SNR, from the coherent ML receiver which has the perfect knowledge of the channel.