Robust Multi-Objective Path Planning for Flying Robots under Wind Disturbance

Abstract

This paper proposes a robust multi-objective path planning algorithm for flying robots carrying out complex missions. When a robot is put into the field, the robot is required to perform complex missions such as visiting sequential goals. We specify these missions using a linear temporal logic and search the path to accomplish the mission for flying robots. Since flying robots are more sensitive to air flow than ground robots, we should plan a path more carefully so that the disturbance by airflow does not cause mission failures or collision with obstacles. In addition, it is important to increase energy effectiveness for stability of flying robots. To achieve these purposes, we propose a multi-objective path planning problem which minimizes the mission failure probability and the moving distance while guaranteeing the safety of the robot and mission completion. We introduce a multi-layer path planning algorithm, where the high-level planner guides the low-level planner by generating a discrete path to accomplish the mission and the low-level planner searches the path to optimize multiple objective functions using a sampling-based RRT search tree. The presented low-level planner can improve the Pareto optimality of the trajectory effectively. We analyze the effectiveness theoretically and evaluate the performance by simulations.