变电站巡检机器人越障视觉伺服路径规划  

Visual Servo Path Planning for Obstacle Surmounting of Substation Inspection Robot

作　　者：李葳挺 LI Wei-ting(China Southern Power Grid Technology Co.,Ltd.,Guangzhou 510080 China;Guangdong Provincial Key Laboratory of New Technology for Smart Grid,Guangzhou 510080 China)

机构地区：[1]南方电网电力科技股份有限公司,广东广州510080 [2]广东省智能电网新技术企业重点实验室,广东广州510080

出　　处：《自动化技术与应用》2024年第6期10-14,共5页Techniques of Automation and Applications

基　　金：广东省智能电网新技术企业重点实验室资助项目(2020B1212070025)。

摘　　要：针对巡检机器人越障路径规划准确性较差,难以跨越障碍物、转弯等问题,提出一种变电站巡检机器人越障视觉伺服路径规划方法。基于摄像机成像和四个坐标系之间的关系,构建巡检摄像机成像模型,采集变电站巡检机器人巡检视觉图像,引入轮廓提取法,获取路径上障碍物轮廓信息,计算障碍物凸包面积,准确定位障碍物。利用小波神经网络,构建变电站巡检机器人越障视觉伺服路径规划模型,以此调整步态,完成巡检机器人越障视觉伺服路径规划。实验测试结果表明,所提方法以最快的速度达到期望状态,收敛速度较快,累积误差一直处于稳定状态。Aiming at the problems of poor accuracy of obstacle surmounting path planning of inspection robot,such as difficulty in surmounting obstacles and turning,a visual servo path planning method for obstacle surmounting of inspection robot in substation is proposed.Based on the relationship between the camera imaging and the four coordinate systems,the inspection camera imaging model is constructed,the inspection visual image of the substation inspection robot is collected,and the contour extraction method is introduced to obtain the contour information of the obstacles on the path,calculate the convex hull area of the obstacles,and accurately locate the obstacles.By using wavelet neural network,a visual servo path planning model of substation inspection robot is constructed to adjust the gait and complete the visual servo path planning of the inspection robot.The experimental results show that the proposed method can reach the desired state at the fastest speed,and the convergence speed is fast,and the accumulated error is always stable.

关 键 词：变电站 巡检机器人 越障路径 伺服视觉规划 路径规划 

分 类 号：TP18[自动化与计算机技术—控制理论与控制工程] TP242.2[自动化与计算机技术—控制科学与工程]