Improved Q-learning based route planning method for UAVs in unknown environment
Runxia Li, Li Hua Fu, Lingling Wang, Xiaoguang Hu · 2019
Unmanned aerial vehicles (UAVs) have demonstrated their application value in many fields. With the rapidly interest in UAV technology, the body of research on autonomous route planning has enjoyed an accelerating rate of novel proposals and developments. Traditional roadmap-based global route planning algorithms and the local route planning algorithms easily trapped in local optimum are difficult to apply in unknown and complex environments. The reinforcement learning algorithm that perceives information through trial-and-error learning provides a new direction for the route planning of UAV. In this paper, a novel route planning algorithm based on improved Q-learning is proposed for UAV in an unknown environment. The prior knowledge is used to guide the UAV to select the action, and a new convergence method based on single state-action value difference is proposed to speed up the convergence rate. Finally, the simulation experiment results show that the proposed method is more effective than the original method, and the algorithm is feasible and effective for UAV route planning.