.png){kind=icon}
.png){kind=icon}
Return
Optimization-Based Path Planning With Artificial Potential Function
.png){kind=icon}
OA
AI
DOI：10.1109/ACCESS.2025.3597311.png)
.png){kind=icon}
Original request for help.png){kind=icon}
Share.png){kind=icon}
Save
Abstract
En 中文
Path planning remains a critical challenge in autonomous ground vehicle (AGV) systems, particularly in dynamic environments where real-time adaptation is essential. Unlike existing methods that often struggle with computational efficiency or environmental adaptability, this study introduces a novel optimization-based approach that seamlessly integrates artificial potential functions with the Levenberg-Marquardt optimization method. The proposed algorithm addresses the key limitations of traditional path planning by simultaneously handling static obstacles and dynamic environmental changes through a unified cost function framework. The innovation lies in formulating the dynamic environment using traversable and non-traversable regions defined by artificial potential functions a computationally efficient approach for real-time obstacle representation. The Levenberg-Marquardt method is specifically chosen for its superior convergence properties and robustness in nonlinear optimization problems compared to gradient-based alternatives. The optimization problem incorporates vehicle dynamics, control constraints, and environmental variations into a single framework. An integral controller ensures a precise path following the generated optimal trajectory. Experimental validation in both simulation and real-world scenarios demonstrates the algorithm’s effectiveness and practical applicability. A validation video is available at https://youtu.be/XVLes855hSw
Keywords:
Autonomous driving system
path planning
model predictive control
optimization-based path planning
virtual/real environment experiments