A Robust and Efficient Path Planning System Based on the Global Euclidean Distance Fields

Abstract

In this paper, we propose a system for fixed-area fixed-point cruising and unknown-area exploration that combines path planning with ESDF global map construction. To build highly accurate global ESDF maps incrementally with small computational resources, we introduce two independent updating queues for inserting and deleting obstacles separately, as well as Indexing Data Structures and Doubly Linked Lists for map maintenance. An efficient and robust path planning for quadrotor UAVs based on obstacle distance and gradient information included in the global ESDF map. The path planning part includes a global path search module at the front end based on the improved A* algorithm and then a local motion trajectory optimization module at the back end based on the B-Splines. Various simulational experiments are used to validate the feasibility, efficiency, and robustness of the proposed method.