Blind tracking of time-varying MIMO channel with Alamouti scheme in interception context

The context of this paper is interception, thus nothing is known on emission. This paper addresses the problem of time-varying multiple input multiple output channel estimation and tracking with Kalman filter for Alamouti scheme. Generally, Kalman filter must be initialized by training sequences, but in interception systems, such sequences don¿t exist. So, we propose in this paper to use blind source separation to estimate the first transmitted symbols and we use them like pilot symbols for initializing Kalman filter. Performances obtained by using a Kalman filter initialized by blind source separation are better than blind source separation.