汽车工程 ›› 2023, Vol. 45 ›› Issue (10): 1933-1943.doi: 10.19562/j.chinasae.qcgc.2023.10.014

所属专题: 智能网联汽车技术专题-规划&决策2023年

• • 上一篇    下一篇

基于改进冲突搜索的智能车库多AGV路径规划

任明辉1,梁军1(),陈龙1,张春2,王云2   

  1. 1.江苏大学汽车工程研究院,镇江 212013
    2.宝胜系统集成科技股份有限公司,扬州 225800
  • 收稿日期:2023-02-26 修回日期:2023-03-29 出版日期:2023-10-25 发布日期:2023-10-23
  • 通讯作者: 梁军 E-mail:liangjun@ujs.edu.cn
  • 基金资助:
    国家自然科学基金(51108209);宝应县重点研发计划项目(BY201908)

Multi-AGV Path Planning for Intelligent Garage Based on Improved Conflict Search

Minghui Ren1,Jun Liang1(),Long Chen1,Chun Zhang2,Yun Wang2   

  1. 1.Automotive Engineering Research Institute,Jiangsu University,Zhenjiang 212013
    2.Baosheng System Integration Technology Company Limited,Yangzhou 225800
  • Received:2023-02-26 Revised:2023-03-29 Online:2023-10-25 Published:2023-10-23
  • Contact: Jun Liang E-mail:liangjun@ujs.edu.cn

摘要:

智能车库中多自主引导车(automated guided vehicle, AGV)的路径规划问题直接影响存取车的效率和安全性。针对智能车库中AGV的任务执行优先级问题,提出了基于改进冲突搜索的路径规划模型(improved conflict-based search with priority, iCBS-pri),该改进模型主要由任务分配(task allocation, TA)、单AGV路径规划(path planning, PP)、多AGV冲突检测与解决(conflict detection and resolution, CDAR)3个模块组成,TA模块将未分配任务分配给AGV,PP模块通过设置直线惩罚函数,减少路径的转弯次数对AGV运行时间的影响以提高AGV任务完成效率,CDAR模块包括冲突检测(conflict detection,CD)子模块和冲突解决(conflict resolution, CR)子模块,CR子模块针对CD子模块检测出的冲突类型,制定基于备用区域(spare zone, SZ)和旁路规划(bypass, BP)的冲突解决策略,以规划多AGV无冲突路线。仿真实验验证了典型场景下的该模型,结果表明:(1)PP模块所提改进A*相较于传统A*算法在路径长度和拐点数量分别减少8.82%和38.62%;(2)任务分配算法的分配成功率达到100%,任务一致性的概率达88.9%;(3)iCBS-pri算法在任务规划成功率方面比iCBS算法平均提升11.3%,算法平均运行时间提升5.93%,进一步提升了智能车库存取车效率。

关键词: 智能车库, 路径规划, 任务执行优先级, 自主引导车, 冲突

Abstract:

Path planning of multiple Automated Guided Vehicles (AGVs) in intelligent garage directly affects the efficiency and security of the vehicle. For the task execution priority of AGVs in RIG, the Improved Conflict-Based Search with priority (iCBS-pri) path planning model is proposed. The improved model is mainly composed of Task Allocation (TA), single-AGV Path Planning (PP), multi-AGV Conflict Detection and Resolution modules. The TA module allocates unassigned tasks to AGVs. The PP module improves the completion efficiency of AGV tasks by setting a linear penalty function to reduce the impact of the number of turns of the path on AGV running time. The CDAR module includes Conflict Detection (CD) submodule and Conflict Resolution (CR) submodule. The CR submodule develops conflict resolution policies based on Spare Zone (SZ) and Bypass planning (BP) for the conflict types detected by the CD submodule, so as to plan multi-AGV conflict-free routes. Simulation experiments verify the model under typical scenarios. The results show that: (1) Compared with the traditional A* algorithm, the improved A* proposed by the PP module reduces the path length and the number of inflection points by 8.82% and 38.62%, respectively; (2) The assignment success rate of the task allocation algorithm reaches 100%, with the task consistency probability reaching 88.9%; (3) Compared with the iCBS algorithm, the success rate of task planning of iCBS-pri algorithm is improved by 11.3% on average, with the average running time of the algorithm improved by 5.93%, which further improves the efficiency of RIG access vehicle.

Key words: intelligent garage, path planning, task execution priority, autonomous guide vehicle, conflict