变电站二次设备巡检的Mecanum轮式机器人精确定位与作业辨识

通过巡检机器人进行变电站二次设备的监测是提升电力设备自动化、智能化管理的重要方式，有利于保障电力工程设备的安全运行。本研究开发了一种用于变电站二次设备自动化巡检的Mecanum轮式机器人，具备自主导航定位与作业辨识的能力，可极大提升设备巡检效率及保护压板状态识别准确性。通过Mecanum轮的驱动方式实现巡检机器人在狭窄作业环境下的灵活运动与姿态调整，多轨道升降平台实现对350-1800 mm高度范围内的二次设备及压板的图像采集与状态辨识。机器人采用基于激光雷达的SLAM导航方法进行自主定位导航，并结合基于视觉的路径提取与跟踪算法进行姿态位置修正，实现在待测点位置的精确定位。同时，提出了基于颜色辨识的图像排列与状态辨识方法，针对二次设备保护压板连通状态进行识别和判断。实验结果表明，研制的变电站二次设备巡检机器人可以实现自主导航与位置精确定位，在路径跟踪过程中最大偏角和偏距分别为±3°和±8 mm。结合机器视觉与颜色辨识的压板辨识方法可以准确识别压板状态，识别准确率大于95.80%，有助于提升机器人自动化的电力巡检作业水平。

Precise positioning and identification of substation secondary equipment inspection robot based on Mecanum wheel

The monitoring of substation secondary equipment through inspection robots is an important way to improve the automation and intelligent management of power equipment, which is conducive to ensuring the safe operation of power engineering equipment. This study develops a Mecanum wheeled robot for automatic inspection of sec-ondary equipment in substations. It has the ability of autonomous navigation, positioning and identification, which can greatly improve the efficiency of equipment inspection and the accuracy of identification of the pro-tection platen state. The flexible movement and attitude adjustment of the inspection robot in a narrow working environment is realized through the driving mode of the Mecanum wheels, and the multi-track lifting platform realizes the image acquisition and state identification of the secondary equipment pressure plate within the height range of 350-1800 mm. The robot adopts the LIDAR-based SLAM method for autonomous positioning and navigation, and combines the vision-based path extraction and tracking algorithm for position correction to achieve precise positioning at the setpoints. At the same time, an image arrangement and state recognition method based on color recognition is proposed to identify and judge the connection state of the secondary equipment protection plate. The experimental results indicate that the substation secondary equipment inspection robot based on the Mecanum wheels can realize autonomous navigation and precise positioning, and the maximum deflection angle and distance during the path tracking process are ±3° and ±8 mm, respectively. The platen recognition method combined with machine vision and color recognition can accurately identify the state of the platen, and the recognition accuracy rate is greater than 95.80%, which helps to improve the level of robot auto-mation inspection operations.