Advanced artificial potential field for safe and efficient local path planning

Purpose This paper addresses the problem of efficient and safe local path planning for mobile robots operating in environments with unknown and randomly distributed obstacles. The purpose of this study is to develop a computationally lightweight local path planner that is designed to support obstacle avoidance, smooth goal-reaching behaviour and reliable waypoint tracking in cluttered environments. Design/methodology/approach The proposed method is developed on the basis of the artificial potential field algorithm and extends the classical formulation by incorporating future position prediction, vortex potential, a single constrained quadratic optimization problem for control signal selection and a mechanism for maintaining a safe distance from obstacles. The algorithm is evaluated in Matlab using randomly generated obstacle configurations for both the standalone local path planning problem and the waypoint tracking problem. Its performance is compared with that of selected existing local path planning methods. In addition, the proposed approach is validated experimentally using a real mobile robot. Findings The obtained results indicate that the proposed local path planner achieves low goal-reaching time, short path length, satisfactory path smoothness and accurate waypoint tracking. The method maintained repeatable obstacle-distance behaviour in the evaluated scenarios while requiring relatively low computational effort. The AAPF guidance field is used to generate a predicted trajectory, which is then checked against the proposed safety-distance criterion. Originality/value The main contribution of this work is the development of an efficient local path planning algorithm that combines safety-distance-aware obstacle avoidance, prediction-based motion planning and constrained optimization-based control signal generation within a unified framework. The proposed method offers a favourable balance between navigation performance, safety and computational complexity, which makes it suitable for real-time implementation in mobile robotic systems.