UAV Swarm Communication Aware Formation Control via Deep Q Network

Abstract

We propose a DQN based reinforcement learning method in swarm communication constraint based formation control with target searching function. A decentralized communication performance indicator is applied in evaluating the UAV’s formation control in simulating a more realistic wireless communication environment between each UAV. A target searching model based on communication aware formation control is presented. The simulation results show that the trained model with state observation space and one thrust action space could be applied in the larger swarm system’s group formation and target point tracking.