基于ROS的动车组智能巡检平台

Intelligent inspection platform for EMU based on ROS

随着高速铁路的飞速发展,自主检测动车组结构件已成为铁路行业关注和亟待解决的问题。提出一个在动车运用所用机器代替人工智能巡检的方案,执行动车故障检测任务。将移动机器人平台与工业机械手集成在一起,并提出一种与其适配的自主控制系统。该系统由硬件设备层、软件数据处理层和行为规划层组成,硬件设备层使用传感器构建执行智能巡检任务的平台;软件数据处理层分为审议、感知、导航和操纵4个模块,审议模块负责协调系统中的所有行为,感知模块处理感官信息,导航模块控制移动机器人到所需位置,操纵模块控制机械臂移动;行为规划层根据当前与任务相关的情况做出不同的决策。通过SLR算法对激光雷达取得的点云数据,采用双层阈值法处理分析数据后识别障碍物,并将其反馈至该系统。系统根据障碍物信息对移动机器人平台和工业机械手发出指令,依托机器人操作系统使用人工势场方法对机器人运动轨迹进行最优规划后并依托Motoman Package作为基本运动规划驱动程序对机器人进行运动控制,并依托Moveit工具对工业机械手进行动作规划,以便其在动车所环境下有效地检测、识别动车故障。通过在长沙城际运用所对整个平台进行实测,验证了方法和系统的可行性及有效性。

With the rapid development of high-speed railway, independent detection of EMU structural parts has become a concern and urgent problem to be solved in the railway industry. Therefore, a scheme of replacing artificial intelligence inspection with machine in EMU operation station was proposed, which can perform EMU fault detection task in EMU operation station. The mobile robot platform was integrated with the industrial manipulator, and an adaptive autonomous control system was proposed. The system consisted of hardware device layer, software data processing layer and behavior planning layer. The hardware device layer used sensors to build a platform for performing intelligent inspection tasks. The software data processing layer was divided into four modules: deliberation, perception, navigation and manipulation. The deliberation module was responsible for coordinating all the behaviors in the system. The perception module processed sensory information. The navigation module controlled the mobile robot to the desired position. The behavior planning layer made different decisions according to the current situation related to the task. By using the SLR algorithm to process the point cloud data obtained by laser radar, the double threshold method was used to identify the obstacles and feed them back to the system. The system sent instructions to the mobile robot platform and the industrial manipulator according to the obstacle information. Based on the robot operating system, the artificial potential field method was used to optimize the trajectory of the robot, and the Motoman Package was used as the basic motion planning driver to control the movement of the robot, and the Moveit tool was used to plan the motion of the industrial manipulatorso that it can effectively detect and identify the fault of the motor vehicle in the environment of the motor vehicle. Finally, the whole platform was tested in the Changsha Intercity Application Institute to verify the feasibility and effectiveness of the method and system.