Time-Varying Channel Estimation Using Superimposed Training Based on Self-Interference Suppression

Time-varying channel estimation for single carrier systems using superimposed training is considered. The variation of a wireless channel is well approximated by a discrete prolate spheroidal basis expansion model (DPS-BEM) as basis functions. A periodic training sequence is superimposed at low power on the information sequence as opposed to being time-multiplexed (TM) with it at the transmitter. A two-step approach is adopted, where, in the first step, we estimate the channel by using only the first-order statistics of the received data. The unknown information sequence acts as interference resulting in a poor signal-to-noise ratio (SNR). We then proposed a low complexity receiving self-interference suppression (RSIS) scheme in the second step where the information interference (termed self-interference) is estimated from detected symbols and iteratively suppressed at the receiver to enhance estimation performance. The RSIS scheme avoids the compromise between self-interference elimination and information integrity of data-dependent superimposed training (DDST) scheme in time-varying channel estimation. Simulation results show that the proposed scheme is better than DDST schemes and competitive with the TM training in term of BER.