基于移动机器人的智能车库关键技术综述

Review on Critical Technologies of Robot-based Intelligent Garages

近年来基于移动机器人的智能车库(Robot-based intelligent garage,RIG)以高容量、高效率和高安全性的特点,在缓解城市“停车难”问题彰显卓越功效。RIG作为智能交通系统的重要组成部分,其核心组件是移动机器人(Automated guided vehicle,AGV),而AGV运行控制依赖RIG关键技术。在分析RIG相关文献的基础上,首先根据RIG运行特点,从“布局方案”“调度策略”和“路径规划”三个方面对现有RIG关键技术梳理。然后,系统回顾了RIG关键技术的理论基础、研究方法、应用现状及优缺点,深入探讨了RIG关键技术在运行过程中所发挥的重要作用。最后,针对RIG运行过程中存在的问题,仍从以上三个方面展望了RIG关键技术的发展趋势。需要强调的是:①运用大数据处理技术,挖掘不同时间段停车需求,指导RIG布局自动优化以平衡停车容量和存取效率需要;②云机器人技术在RIG调度策略应用逐渐广泛,结合出行即服务技术RIG,制定满足用户满意度的RIG调度策略;③随着RIG布局方案、调度策略以及实际环境障碍物的复杂化,利用深度强化学习框架解决AGV的路径规划问题已成为当前发展趋势。系对RIG关键技术研究成果的系统梳理,对未来RIG技术研发和工程应用具有重要的指导意义。

In recent years,robot-based intelligent garage(RIG)has shown its effectiveness in alleviating the problem of“difficult parking”in cities with its high capacity,high efficiency,and high safety.The core component of a RIG is an automated guided vehicle(AGV),and the control of AGV operation relies on critical RIG technologies.Based on the analysis of RIG-related literature,the existing RIG technologies are firstly reviewed in terms of“layout scheme”,“scheduling strategy”and“path planning”,based on the operational characteristics of RIGs.The critical component technologies are reviewed.Then,the theoretical basis,research methods,application status,advantages and disadvantages of RIG critical technologies are systematically reviewed,and the important role played by RIG critical technologies in the operation process is thoroughly discussed.Finally,in view of the problems existing in the operation of RIG,the development trend of RIG critical technologies is still foreseen from the above three aspects.It is important to emphasize that①Big data processing technology is used to tap into the parking demand at different time periods and guide the automatic optimization of RIG layout to balance the parking capacity and access efficiency demands;②Cloud robotics is gradually being widely applied in RIG scheduling strategies,which are combined with the mobility-as-a-service technology RIG to develop RIG scheduling strategies that meet user satisfaction;③As RIG layout schemes,scheduling strategies and the complexity of the actual environmental obstacles,the use of deep reinforcement learning framework to solve the path planning problem of AGVs has become the current development trend.The systematization of the research results of critical RIG technologies is an important guide for future RIG technology development and engineering applications.