【摘要】 针对果园非结构化场景中履带机器人全局规划效率低、路径冗余、动态避障性能不足及平滑性差等问题，提出一种融合改进A~*算法、优化人工势场法（Artificial Potential Field,APF）与分段Bezier曲线的三阶段混合路径规划方法。先借助启发函数动态加权与转向惩罚函数优化A~*算法，提升搜索效率并减少冗余节点，再通过优化势场函数模型改进APF，解决局部最小值问题并增强动态避障能力，最后采用分段Bezier曲线完成平滑路径处理。仿真实验表明，该算法在规划效率、路径长度、动态避障能力及平滑度上表现优异，能满足果园履带机器人高效自主作业需求。更多还原

【Abstract】 Aiming at persistent issues of tracked robots in unstructured orchard scenarios-low global planning efficiency, path redundancy, insufficient dynamic obstacle avoidance performance and poor smoothness-a three-stage hybrid path planning approach is proposed, which integrates the improved A~* algorithm, optimized Artificial Potential Field(APF) method and piecewise Bezier curve. The A~* algorithm is first optimized via dynamic weighting of the heuristic function and turning penalty function to enhance search efficiency while reducing redundant nodes, followed by APF improvement through optimizing the potential field function model to resolve the local minimum problem and strengthen dynamic obstacle avoidance capability. Eventually, path smoothing is achieved using the piecewise Bezier curve. Simulation experiments demonstrate that this algorithm performs remarkably well in planning efficiency, path length, dynamic obstacle avoidance and smoothness, fully meeting the requirements of efficient autonomous operation for tracked orchard robots.更多还原