A Multi-Source Potential Field for Off-Road Path Planning

Off-road environments with varied terrain and obstacle types pose challenges for safe maneuvering of unmanned ground vehicles (UGVs). In this paper, the multi-source off-road potential field (MOPF) is proposed to quantify the risk and impedimental effects on vehicle travel in off-road environments b...

Full description

Saved in:
Bibliographic Details
Published in2023 5th International Conference on Robotics, Intelligent Control and Artificial Intelligence (RICAI) pp. 416 - 422
Main Authors Xie, Yujia, Nie, Shida, Liu, Hui, Han, Lijin, Guo, Congshuai, Zhang, Fawang
Format Conference Proceeding
LanguageEnglish
Published IEEE 01.12.2023
Subjects
3D terrains
Artificial intelligence
artificial potential field
autonomous vehicles
Collision avoidance
Land vehicles
Path planning
Robots
Safety
Vehicle dynamics
Online AccessGet full text

Cover

More Information
Summary:Off-road environments with varied terrain and obstacle types pose challenges for safe maneuvering of unmanned ground vehicles (UGVs). In this paper, the multi-source off-road potential field (MOPF) is proposed to quantify the risk and impedimental effects on vehicle travel in off-road environments by taking the effect of the three-dimensional terrain and obstacles into account in combination with the vehicle performance and state. First, vehicle failure conditions are defined based on dynamics analysis, and crossable and non-crossable terrain are classified to consider the limitations of vehicle performance on various terrain. The non-uniform safety margin potential field (NSMPF) is proposed, based on which the non-crossable terrain source potential field (NTSPF) and the non-crossable obstacle source potential field (NOSPF) are established to quantify the risk, which can help vehicle effectively avoid unnecessary detours and save energy. The crossable obstacle source potential field (COSPF) and terrain source potential field (TSPF) are established to represent the impedimental effect of off-road environment on vehicle traveling. Finally, an application example of the MOPF in path planning is given. The MOPF has been proved by simulation to generate paths with short lengths and match the vehicle performance, while ensuring collision avoidance. In addition, the MOPF can also include different human needs and driving styles.
DOI:10.1109/RICAI60863.2023.10489546