Integrated Sensing and Communication for UAV Detection: A Reinforcement Learning-based Approach

The detection efficiency for unmanned aerial vehi-cles (UAV s) is lower as the targets move at low altitude and low velocity. In this paper, we propose a detection method for low-altitude small UAV s using the integrated sensing and communication (ISAC) in 5G networks. We first present a reflection wave model for UAV s in multipath environments covered by 5G networks. Then, we propose an adaptive detector based on reinforcement learning. The detector uses a reward mechanism based on the minimum signal-to-noise ratio (SNR) to construct a reinforcement learning model, and performs round-based iterative update of the parameter selection policy independent of human experience. Experiments demonstrate that the success rate of target detection of our proposed low-altitude UAV detection can achieve 91.2 %.

I. Introduction

With the rapid development of unmanned aerial vehicles (UAV s), the characteristics of low flight height, low flight velocity, and small reflection area lead to the difficulty for target detection. Therefore, how to improve the capability of detecting small low-altitude airspace targets has become a challenge that needs to be urgently addressed in the air traffic control field in recent years [11], [21].