A Comprehensive Review of Artificial Potential Field Techniques in Human–Robot Collaboration from 2019 to 2024

With the increasing application of collaborative robots (cobots) in manufacturing and service sectors, ensuring smooth motion planning and collision-free interaction is a major concern in Human–Robot Collaboration (HRC). Among different local path planning schemes, the Artificial Potential Field (APF) approach has become one of the most popular strategies due to its simplicity and effective computation. However, traditional APF models still have some challenges, such as local minima, limited adaptability in dynamic environments, and no integration with human-aware perception. This study delivers a systematic review of the related developments of APF approaches in HRC for the period of 2019-2024. Totally, 169 papers were first collected from major scientific databases, and then 32 peer-reviewed articles with strict selection criteria were chosen for in-depth analysis. This work is divided into three broad research avenues: (i) algorithmic improvement of APF towards better stability and adaptability, (ii) integration of sensors and Machine Learning (ML) to provide smart perception, and (iii) HRC for improvement of interaction quality and security. The results of the review suggest an intense trend towards hybrid structures integrating APF with Reinforcement Learning (RL), biosensing, and Digital Twin (DT) technologies. These integrations significantly enhance real-time responsiveness, safety, and robustness in dynamic conditions. Finally, this review highlights several limitations and outlines further research directions for improving APF-based methods for more perceptive and human-centered robot systems.