基于Dubins曲线和遗传算法的AUV路径规划方法  

An AUV Path Planning Method Based on Dubins Curves and Genetic Algorithm

作　　者：张骁 韩云涛[1] 许振 何昱霖 ZHANG Xiao;HAN Yun-tao;XU Zhen;HE Yu-lin(College of Automation,Harbin Engineering University,Harbin 150001 China;China National Aeronautical Radio Electronics Research Institute,Shanghai 200233 China)

机构地区：[1]哈尔滨工程大学自动化学院,黑龙江哈尔滨150001 [2]中国航空无线电电子研究所,上海200233

出　　处：《自动化技术与应用》2020年第12期7-10,100,共5页Techniques of Automation and Applications

基　　金：国家自然科学基金项目(编号51209049);黑龙江省自然科学基金项目(编号E2017015);哈尔滨工程大学教改项目(编号JG2019B14)。

摘　　要：针对遗传算法在传统的网格地图上搜索路径不平滑及不能满足无缆水下机器人(AUV)自身运动学约束问题,提出一种基于Dubins曲线和遗传算法的AUV路径规划方法。在运动约束条件下,利用平面直角坐标系进行环境建模,将障碍物等效成圆形,提出检测路径安全性的检测算子,并设计Dubins曲线的三进制编码方案生成初代染色体,再利用遗传算法,通过选择、交叉、变异、解码、检测等操作得到最终的路径。通过对多种情况下不同半径的障碍物进行仿真,结果表明该方法可以搜寻到避开障碍物且满足运动学约束的相对较短的路径。To solve the problem that the search path of genetic algorithm on traditional grid map is not smooth and can not satisfy the kinematic constraint of Autonomous Underwater Vehicle(AUV).An AUV path planning method based on Dubins curves and genetic algorithm is proposed.Under the condition of motion constraints,build the environment modeling in rectangular coordinates,turning obstacles into circles,proposing a detection operator for path security and the ternary coding scheme with Dubins curves is designed to generate primary chromosomes.Then using Genetic algorithm obtains the final path through selection,crossover,mutation,decoding and detection.Through the simulation of obstacles with different radii under various conditions.The results show that the method can find a relatively short path that avoids obstacles and satisfies kinematic constraints.

关 键 词：Dubins曲线 遗传算法 路径规划 无缆水下机器人 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP242[自动化与计算机技术—控制科学与工程]