An intelligent distributed anti-collision control algorithm for UAV swarm formation

Abstract:
To improve the control quality and anti-collision performance of the UAV swarm, this paper proposes a new intelligent anti-collision swarm control algorithm based on the advantages of two algorithms, consistency algorithms and reinforcement learning algorithms. The consistency algorithm performs less well in anti-collision aspect, and the reinforcement learning algorithm is unable to convergence in finite time. For eliminating adverse effects of communication jamming environments, this paper combines the consistency algorithm with Fuzzy Q learning algorithm, presenting a new Multi- Agent Consistency-Fuzzy Q learning algorithm (MACFQL). The validity of the algorithm is demonstrated in the section of calculating samples.