Signal Detection for Ambient Backscatter Communications Based on DDPG

In this paper, a modified model of the ambient backscatter communication (AmBC) using reconfigurable intelligent surface (RIS) is built. On this basis, the optimization problem of solving the maximum value of generalized relative channel difference for signal detection is derived. The optimization solution based on deep deterministic policy gradient (DDPG) is proposed. Monte-Carlo simulation results show that the proposed optimization method has the advantages of high gains and fast convergence rate.