基于视觉SLAM算法的轮式机器人位置误差标定  

Position error calibration of wheeled robots based on visual SLAM algorithm

作　　者：耿中宝 宋亚磊 GENG Zhongbao;SONG Yalei(College of Information Mechanical&Electrical Engineering,Zhengzhou Business University,Zhengzhou 451200,China;School of Computer Science and Technology,Zhengzhou University of Light Industry,Zhengzhou 450000,China)

机构地区：[1]郑州商学院信息与机电工程学院,郑州451200 [2]郑州轻工业大学计算机与通信工程学院,郑州450000

出　　处：《兵器装备工程学报》2024年第12期306-312,共7页Journal of Ordnance Equipment Engineering

基　　金：河南省本科高校智慧教学专项研究项目(78);河南省高等学校重点科研项目(24B790031);河南省科技攻关项目(242102210147)。

摘　　要：现有方法在运动位置误差标定时仅以里程误差作为位置标定参数,难以满足大空间范围、高精度的测量需求。因此,提出基于视觉SLAM算法的轮式机器人位置误差标定方法。采用Kinect相机采集轮式机器人的运动环境数据,再利用视觉SLAM算法建立机器人的环境地图,以适应大空间范围的测量需求,提高位置误差标定的精度和可靠性。对里程误差及视觉采集误差实施标定迭代,结合两者标定结果,精确标定轮式机器人的位置误差。结果表明:在直线轨迹与曲线轨迹下,所提方法控制下位置误差最大不超过10 mm,可降低轮式机器人的位置标定误差,提高运动轨迹的控制精度。The existing methods only use mileage error as a position calibration parameter in motion position error calibration,which is difficult to meet the measurement requirements of large spatial range and high accuracy.Therefore,a position error calibration method for wheeled robots based on visual SLAM algorithm is proposed.This paper uses Kinect cameras to collect motion environment data of wheeled robots,and then uses visual SLAM algorithm to establish the robot’s environment map,in order to adapt to the measurement needs of a large spatial range and improve the accuracy and reliability of position error calibration.This paper implements calibration iteration for mileage error and visual acquisition error,and combines the calibration results of both to accurately calibrate the position error of the wheeled robot.The results show that under the control of the proposed method,the maximum position error does not exceed 10 mm in both linear and curved trajectories,which can reduce the position calibration error of wheeled robots and improve the control accuracy of motion trajectories.

关 键 词：视觉SLAM算法 Kinect相机 轮式机器人 里程误差 误差标定 

分 类 号：TP242[自动化与计算机技术—检测技术与自动化装置]