基于改进D^(*)算法的AUV三维动态路径规划  

Three-dimensional Dynamic Path Planning of AUV Based on Improved D^(*) Algorithm

作　　者：朱蟋蟋 孙兵[1] 朱大奇[1] ZHU Xi-xi;SUN Bing;ZHU Da-qi(Shanghai Engineering Research Center of Intelligent Maritime Search&Rescue and Underwater Vehicles,Shanghai Maritime University,Shanghai 201306,China)

机构地区：[1]上海海事大学智能海事搜救与水下机器人上海工程技术研究中心,上海201306

出　　处：《控制工程》2021年第4期736-743,共8页Control Engineering of China

基　　金：国家自然科学基金资助项目(61873161,U1706224);上海市青年科技启明星计划资助项目(20QA1404200);国家重点研发计划课题(2017YFC0306302)。

摘　　要：传统D^(*)算法应用于路径规划时,规划的路径一般由折线段组成,且紧靠障碍物边缘。由于自身运动学约束,自治水下机器人(autonomous underwater vehicle, AUV)难以实现折线等情况的快速转向;同时由于自身体积约束,AUV可能会碰撞到障碍物边缘。针对AUV三维水下动态路径规划问题,在原D^(*)算法的代价函数基础上增加了障碍物威胁代价项,以保证规划的路径更安全;利用均匀B样条曲线拟合模型进行路径平滑处理,进而得到优化的AUV规划路径。最后,仿真结果表明,改进的D^(*)算法可以达到预期效果。When the traditional D^(*) algorithm is applied to path planning, the planned path is generally composed of fold lines and close to the edges of obstacles. Due to its own kinematic constraints, it is difficult for autonomous underwater vehicle(AUV) to achieve rapid steering of the fold line. At the same time, due to its own volume constraints, AUV may collide with the edges of obstacles. Aiming at the three-dimensional underwater dynamic path planning problem of AUV, the obstacle threat cost item is added based on the cost function of the original D^(*) algorithm to ensure that the planned path is safer. The uniform B-spline curve fitting model is used to smooth the path, and then an optimized AUV planning path is obtained. Finally, the simulation results show that the improved D^(*) algorithm can achieve the expected results.

关 键 词：自治水下机器人 三维路径规划 D^(*)算法 威胁代价 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]