Motion Planning for Autonomous Ground Vehicles Using Artificial Potential Fields: A Review

TL;DR

This paper reviews various artificial potential field algorithms for motion planning in autonomous ground vehicles, analyzing their improvements, applications, and performance in static and dynamic environments.

Contribution

It provides a comprehensive survey of modified APF algorithms, evaluating their effectiveness and proposing future research directions.

Findings

Modified APF algorithms address traditional shortcomings.

Performance varies based on environment and scenario.

Future research should focus on real-time dynamic adaptation.

Abstract

Autonomous ground vehicle systems have found extensive potential and practical applications in the modern world. The development of an autonomous ground vehicle poses a significant challenge, particularly in identifying the best path plan, based on defined performance metrics such as safety margin, shortest time, and energy consumption. Various techniques for motion planning have been proposed by researchers, one of which is the use of artificial potential fields. Several authors in the past two decades have proposed various modified versions of the artificial potential field algorithms. The variations of the traditional APF approach have given an answer to prior shortcomings. This gives potential rise to a strategic survey on the improved versions of this algorithm. This study presents a review of motion planning for autonomous ground vehicles using artificial potential fields. Each article is evaluated based on criteria that involve the environment type, which may be either static or dynamic, the evaluation scenario, which may be real-time or simulated, and the method used for improving the search performance of the algorithm. All the customized designs of planning models are analyzed and evaluated. At the end, the results of the review are discussed, and future works are proposed.