Blind Channel Estimation and Symbol Recovery Based on DDPG Algorithm

Blind channel estimation (BCE) plays a critical role in wireless systems, aiming to recover channel information without relying on pilot signals. To enable accurate joint recovery of transmitted symbols and estimation of multipath channels, this work proposes a novel online algorithm using deep reinforcement learning (DRL), specifically the Deep Deterministic Policy Gradient (DDPG) framework. The method directly outputs equalized symbols, with the multipath channel response modeled as trainable parameters. Simulation results show that the proposed algorithm outperforms advanced methods in bit error rate (BER) and channel estimation mean squared error (MSE), and is highly competitive with conventional pilot-based methods.