Doubly Selective Channel Estimation and Symbol Detection with Data-dependent Superimposed Training for OFDM Systems

摘要

A channel estimation using Data-dependent Superimposed Training (DDST) over doubly selective channel is proposed for OFDM systems. Instead of using traditional pilot-assisted scheme, this scheme uses training superimposed to data sequence in time domain to assist the estimation, which consumes no extra bandwidth. The superimposed training is designed data-dependent to eliminate the interference of unknown data on estimation thoroughly. Meanwhile, an iterative symbol detection method is presented for the DDST based time varying channel. Simulations show that the proposed estimator can approach the real channel more accurately and achieve better Bit Error Ratio (BER) performance.