Chen, Long (Concordia University), Mantegh, Iraj (National Research Council Canada), He, Tong (Concordia University), XIE, WENFANG (Concordia University)

Fuzzy Kinodynamic RRT: A Dynamic Path Planning and Obstacle Avoidance Method

Scheduled for presentation during the Regular Session "Autonomy II" (WeB1), Wednesday, September 2, 2020, 16:00−16:20, Macedonia Hall

2020 International Conference on Unmanned Aircraft Systems (ICUAS), September 1-4, 2020 (Postponed from June 9-12, 2020), Athens, Greece

This information is tentative and subject to change. Compiled on April 20, 2024

Abstract

Path planning is the essential capability for autonomous navigation of UAV (Unmanned Aerial Vehicle) in unknown environments. In this paper, a Fuzzy logic inferencing system has been designed to achieve obstacle avoidance in a dynamic environment. We introduce Fuzzy-Kinodynamic RRT, method which generates dynamic path based on the traditional rapidly exploring random tree (RRT) algorithm. A set of simple Fuzzy rules are proposed for simple 2D and 3D path planning cases. It is an optimized path planning method which uses Kinodynamic RRT algorithm to do global path planning and utilizes Fuzzy logic to avoid obstacles. A set of heuristics Fuzzy rules are proposed to lead the UAV away from un-modeled ground-based obstacles and to guide the UAV towards the goal. In addition, the designed Fuzzy rules can augment traditional RRT for dealing with new obstacles in the environment. Various simulations are conducted in 2D and 3D environment and the results illustrate the effectiveness of the algorithm in simple dynamic environment.