Low-Complexity Discriminatory Channel Estimation Based on Training Sequence Contamination

Channel estimation security is an important aspect of physical layer security and training sequences play a key role in channel estimation. The method proposed in this letter employs CSI and CFO as the contamination information of training sequences. Legitimate transceivers contaminate and decontaminate training sequences through low-complexity mathematical operations, significantly discriminating the channel estimation results of legitimate receiver and eavesdropper. The simulation results indicate that compared with the traditional AN, MFB-DNN and LS-DCE methods, the proposed method has a more significant differentiation effect, and its computation complexity is much lower than that of AN and MFB-DNN methods.