3D Path Planning Based on Dynamic Goal Bias and Critical Region Sampling

This paper presents a path planning algorithm based on dynamic target bias (Bi), an improved artificial potential field (APF) model, and a critical region sampling strategy, referred to as the Bi-APF-RRT* (dynamically-target-biased artificial potential field and rapid random tree) algorithm, aimed at optimizing path planning in complex 3D environments. Experimental results show that the proposed algorithm significantly outperforms traditional RRT* and APF-RRT* algorithms in terms of both path length and smoothness. The dynamic target bias helps mitigate the local optimization problem inherent in traditional methods, while the improved APF model enhances obstacle avoidance and reduces path oscillations. The critical region sampling strategy improves the algorithm's adaptability and robustness, making these improvements particularly suitable for complex environments, such as autonomous driving.