1.重庆理工大学 计算机科学与工程学院;2.重庆大学 自动化学院
V249.32
中国重庆市自然科学基金（批准号：cstc2018jcyjAX0835）
1.Chongqing University of Technology;2.Chongqing University College of Automation
Natural Science Foundation of Chongqing Municipality in China (Grant numbers cstc2018jcyjAX0835)
针对多障碍物场景下移动机器人路径规划存在无解、转向速度状态变化大和路径与障碍物距离过近问题，提出了一种基于代价函数改进的车辆约束路径规划算法。该方法以车辆约束路径规划算法Hybrid A*为基础算法，以其代价函数为原始代价函数。建立双起点和双终点的搜索过程，形成双向搜索，降低路径搜索时间。引入转向约束，降低传统Hybrid A*算法的转向速度状态变化。引入距离约束，降低危险障碍边缘的节点的优先级。实验证明，相较于传统车辆约束Hybrid A*算法，改进的Hybrid A*算法平均路径搜索时间降低了12.043%，转向速度状态变化降低了16.623%，与危险障碍边缘相近的节点优先级降低了25%。这一系列实验结果成功改善了传统Hybrid A*算法在规划路径算法，为多障碍物场景下移动机器人路径规划提供了有效的解决方案。
Aiming at the problems of unsolved mobile robot path planning in multi-obstacle scenarios, large change of steering speed state and too close distance between path and obstacles, a vehicle constrained path planning algorithm based on the improvement of cost function is proposed. The method takes the vehicle constrained path planning algorithm Hybrid A* as the base algorithm and its cost function as the original cost function. A dual start and dual end search process is established to form a two-way search and reduce the path search time. Introduce steering constraints to reduce the steering speed state change of the traditional Hybrid A* algorithm. Introduce distance constraints to reduce the priority of dangerous edge nodes. The experiments demonstrate that compared to the traditional vehicle constraint Hybrid A* algorithm, the improved Hybrid A* algorithm reduces the average path search time by 12.043%, the steering speed state change by 16.623%, and the priority of nodes close to the danger edge by 25%. This series of experimental results successfully improves the traditional Hybrid A* algorithm in planning path algorithm and provides an effective solution for mobile robot path planning in multi-obstacle scenarios..
