Automatic Gain Control of Wireless Receiver Based on Q-Learning

Abstract

Abstract—In the wireless communication system, due to the complexity of the physical channel, the amplitude of the signal received by the wireless receiver often fluctuates wildly, which will increase the bit error rate of signal demodulation. Therefore, the automatic gain control (AGC) is an essential part of the wireless receiver, which can adaptively adjust the gain of each part of the receiver and provide a stable input for the subsequent circuit. Artificial intelligence technology has developed, and reinforcement learning in signal processing has received extensive attention. This paper proposes a gain automatic control method based on Q-learning in the zero-IF receiver, which uses the Q-learning model to learn the characteristics of signal amplitude changes to adjust the speed of the gain adjustment and to track the signal changes more accurately. The simulation results show that the AGC proposed in this paper is more stable than the traditional AGC without Q-learning and can quickly compensate for significant changes in Orthogonal Frequency Division Multiplexing (OFDM) signals.