Semiblind tracking and equalization of time-varying multipath channels

We propose a Kalman filtering based semiblind approach to tracking and equalizing time-varying multipath channels encountered in mobile communications. By using a linear precoder at the transmitter to introduce a priori known correlation into the transmitted signals, the receiver can estimate the channel second order statistics (SOS) blindly. This allows the proposed channel tracking method to be able to initialize with very short (as short as the channel order) training sequences, as compared to existing methods where frequent and long training is needed. We establish the channel SOS identifiability condition and the design constraints on the linear precoders. Linear precoders that satisfy the constraints and achieve diversities are constructed. Simulation results show the effectiveness of the proposed method.