Author:
Zheng Deyan,Liu Chunhui,Huang Lizhen
Abstract
Abstract
A spatio-temporal coverage planning algorithm for multi-UAV and multi-sensor cooperative search is proposed in this paper, named STC-MARL, which is based on multi-agent reinforcement learning and allows several UAVs equipped with two types of sensors to learn to complete full view of a field of interest (FOI). The proposed STC-MARL algorithm can make participating UAVs to learn from the environment to complete the full coverage of a specific FOI while minimizing field of views (FOVs) interacted with each other. The experimental results show in detail with simulation that the UAVs in the mission can effectively accomplish the task of spatio-temporal coverage planning.
Subject
General Physics and Astronomy
Reference8 articles.
1. Collaborative Coverage Path Planning of UAV Cluster based on Deep Reinforcement Learning;Dong;IEEE International Conference on Frontiers Technology of Information and Computer,2021
2. Obstacle avoidance in a dynamic environment: a collision cone approach;Chakravarthy;IEEE Transactions on Systems Man & Cybernetics,1998
3. Deep Reinforcement Learning Based Collaborative Optimization of Communication Resource and Route for UAV Cluster;Huang,2021
4. A reinforcement learning algorithm in cooperative multirobot domains;Borrajo;Journal of Intelligent and Robotic Systems,2005