Fractional-Order Artificial Potential Field with Global Guidance for Bidirectional and Adaptive Path Planning in Mobile Robots

To address the limitations of the traditional A^∗ algorithm, including excessive node expansion and non-smooth paths, this paper proposes a bidirectional dynamic A^∗ algorithm guided by an improved fractional-order artificial potential field (FO-APF). The core innovation involves reformulating the potential field using fractionalorder calculus to generate a superior guiding force that dynamically constrains the search neighborhood, thereby minimizing redundant expansions. The algorithm further integrates adaptive heuristic weights and a bidirectional strategy to accelerate convergence, with subsequent path smoothing achieved via Bézier curve fitting. MATLAB simulations on 30 × 30 and 50 × 50 maps demonstrate significant reductions in search time (up to 47.31%), expanded nodes (33.46%), and number of turns. These results confirm the method’s efficacy in enhancing path planning efficiency and quality, particularly in complex environments.