基于多智能体强化学习的道路施工自动化监控系统

Road construction automation monitoring system based on multi⁃agent reinforcement learning

针对全方位监控复杂多变的道路施工环境,提高施工质量与效率的目的,采用多智能体强化学习算法设计道路施工自动化监控系统。通过信息采集智能体采集道路施工现场环境信息,传输采集的道路施工现场环境信息至现场监控智能体,现场监控智能体依据接收的环境信息,利用多智能体强化学习技术,制定合理的道路施工调度决策,并下发至施工环境智能体,施工环境智能体根据接收的调度指令执行道路施工任务,完成布置的指定施工任务,通过远程监控智能体远端保存与可视化展示道路施工相关数据。实验证明,文中系统可有效采集道路施工现场的风速环境信息,风速主要集中在3 m/s至7 m/s之间,制定道路施工调度决策,提升道路施工质量。

In order to comprehensively monitor the complex and changeable road construction environment and improve the construction quality and efficiency,the multi-agent reinforcement learning algorithm is used to design the road construction automation monitoring system.The information acquisition agent collects the environmental information of road construction site and transmits it to the on-site monitoring agent.According to the received environmental information,the on-site monitoring agent uses multi-agent reinforcement learning technology to make a reasonable road construction scheduling decision,which is sent to the construction ambient intelligence.The construction ambient intelligence performs the road construction task according to the received scheduling instruction and completes the assigned construction task.The remote monitoring agent saves and visually displays the road construction related data.Experiments show that this system can effectively collect the information of wind speed environment on the road construction site,and the wind speed is mainly between 3 m/s and 7 m/s,so as to make the decision of road construction scheduling and improve the quality of road construction.