Time-selective MIMO channel estimation based on implicit training

A novel scheme to perform multiple input multiple output (MIMO) channel estimation in time-selective fading environments is suggested. In order to improve transmission efficiency, training sequences are arithmetically added to the information symbols. We regard the output of information symbols, additive white noises and dc-offsets at the receiver as fiction measurement noise, whose mean and variance are both unknown. A time-varying measurement noise recursive estimator and Kalman filters cooperate to track the time-varying MIMO channel impulse response (CIR). Finally,it is compared with some existing methods, and all these indicate that the proposal exhibits better performance.