An Improved Artificial Potential Field Algorithm Based on a Guided Path Construction Mechanism

This paper proposes an improved Artificial Potential Field (APF) algorithm that integrates a guided path construction mechanism with an ellipse-assisted potential field to address the limitations of traditional APF in complex environments. The method first employs a segment-wise parametric modeling strategy to generate a smooth and continuous global reference path, thereby avoiding the discontinuities and sharp turns common in sampling- or grid-based approaches. On this basis, a path-attraction force and an auxiliary disturbance force are incorporated into the APF model to enhance trajectory adherence and enable the robot to escape from local minima. The resulting composite potential field effectively combines target attraction, obstacle repulsion, path guidance, and elliptical perturbation, significantly improving robustness, smoothness, and real-time performance in dynamic and cluttered environments. Simulation analyses demonstrate that the proposed approach maintains low computational complexity while providing high reliability and strong adaptability for autonomous mobile robots in semi-structured scenarios.