Cooperative System of Ground Robots based on dynamic exploration of Drone Occupancy Grid Maps

摘要

This study proposes a real-time action planning system for autonomous mobile robots, allowing them to request assistance from other robots when necessary. The system focuses on search and uses a ground robot and a drone to search unknown areas. The drone creates an Occupancy Gird Map and transfers it to the ground robot for optimal path planning. The Behavior Trees used for autonomous robot search reactively select actions to uncertain events by monitoring the state, which is grown using a reachability graph generated in advance by offline planning. The system allows robots to automatically generate tasks and decide which robot should perform them, optimizing battery consumption and avoiding unnecessary movements.