智能仓库中基于两层架构的群机器人路径规划  

Path Planning of Swarm Robots Based on Two-tier Architecture in Intelligent Warehouse

作　　者：薛颂东[1] 张宇[1] 赵静 潘理虎[1] XUE Songdong;ZHANG Yu;ZHAO Jing;PAN Lihu(Taiyuan University of Science and Technology,Taiyuan 030024;Guangdong Mechanical&Electrical Polytechnic,Guangzhou 510515)

机构地区：[1]太原科技大学,太原030024 [2]广东机电职业技术学院,广州510515

出　　处：《计算机与数字工程》2023年第9期2026-2032,2108,共8页Computer & Digital Engineering

基　　金：山西省软科学项目(编号:2019041010-2);山西省高校教学改革创新项目(编号:J2019133);山西省哲社科学规划课题一般课题(编号:2020-270);广东省普通高校特色创新项目(编号:2018GKTSCX057);广东省普通高校重点科研平台和项目(编号:2020KCXTD067)资助。

摘　　要：为提高智能仓库中群机器人的拣选作业效率,提出基于集成K-means聚类算法、遗传算法和改进A*算法的群机器人两层多目标路径规划算法。建立仓库模型后,设计交通规则;在任务层,用融合K-means聚类的改进遗传算法进行任务分配;在行为层,以最短运行时间为目标,用考虑转弯代价的改进A*算法进行运动控制,通过启发函数将搜索方向指向目标点,计算成员机器人的多目标节点路径矩阵;在底层,用预约表化解机器人运行过程中的局部冲突。最后,进行仿真实验并与现有路径规划方法进行比较。结果表明,论文方法能有效缩短群机器人的总运行路程和时间。To improve the efficiency of picking operations for swarm robots in intelligent warehouses,an integrated K-means clustering algorithm,genetic algorithm and A*algorithm based path planning algorithm is proposed.First,a warehouse model is built and the traffic rules are designed.Then an improved genetic algorithm fusing K-means clustering algorithm is used for task allocation for each robot at the task level.And then the improved A*algorithm considering turning cost is used for robot moving control by calculating the multi-objective node path matrix of each robot with the shortest moving time as the goal at the behavior level.The search direction is pointed to goal by using the heuristic function.In the moving process,a reservation table is introduced and used for collision avoidance for robots at the bottom level.Finally,a set of experimental simulation are carried out.The results show that the proposed method can dominate the compared existing methods over the total moving distance and the used time.

关 键 词：智能仓库 群机器人 路径规划 群智能优化 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程]