Intelligent Array Antenna in Complex Environment Based on Deep Deterministic Policy Gradient Algorithm

To reduce the loss of phased array antenna gain in complicated electromagnetic environment, a phase regulation algorithm based on deep reinforcement learning (DRL) is investigated in this letter. At first, we analysis and design a DRL-based interactive system between the phased array antenna and inhomogeneous electromagnetic environment. A typical policy-based DRL, namely Deep Deterministic Policy Gradient (DDPG), which has excellent fitting ability for high-dimension continuous action space, is selected to deal with the radiation pattern synthesis problem of the phased array antenna located around metal obstacles. Then, the configuration of a $2 \times 8$ phased array antenna and the radiation environment is given, and the training process of DDPG-based phase regulation algorithm is designed. After completing the training, the phased array antenna with DDPG-based phase regulation algorithm in complicated radiation environment is achieved, which compensates the antenna gain loss by 60% on average.