面向机器人遥操作的数据手套设计及仿真验证  

Design and Simulation Validation of Data Glove for Robot Teleoperation

作　　者：苑朝 王少君 蒋阳 王祝 翟永杰[1] YUAN Chao;WANG Shao-jun;JIANG Yang;WANG Zhu;ZHAI Yong-jie(Department of Automation,North China Electric Power University,Baoding 071000,China)

机构地区：[1]华北电力大学自动化系,河北保定071000

出　　处：《仪表技术与传感器》2022年第9期47-56,共10页Instrument Technique and Sensor

基　　金：国家自然科学基金(61773160);青年科学基金项目(61903033)。

摘　　要：针对面向机器人遥操作的数据手套精度低、成本高以及惯性测量单元校正难、校正不完全等问题,基于9轴惯性测量单元设计了一款数据手套,对传感器数据进行了校正、融合,并在Simulink中搭建了仿真系统。采用LM算法校正加速度计和陀螺仪数据,椭球拟合算法校正磁力计数据;采用扩展卡尔曼滤波算法融合校正后的9轴数据,得到传感器的实时姿态。实验结果表明:经校正、融合得到的传感器三轴动态角度,其均方根误差小于1.5°,最大误差绝对值小于3°,计算得到的手指关节静态角度误差小于1°。该数据手套系统可以较准确地检测手指的姿态,并能实时控制虚拟灵巧手的运动,同时也能获得手指关节运动的角度、角速度等信息。Aiming at the problems of low accuracy and high cost of data gloves for robot teleoperation and difficult and incomplete calibration of inertial measurement unit,a data glove was designed based on 9-axis inertial measurement units.The data of sensors was calibrated and fused.Then a simulation system was built in Simulink.The LM algorithm was used to calibrate the data of accelerometer and gyroscope,and the ellipsoid fitting algorithm was used to calibrate the data of magnetometer.The extended Kalman filter algorithm was used to fuse the calibrated 9-axis data to obtain the real-time attitude of the sensor.The experimental results show that the root mean square error of the three-axis dynamic angle of the sensor is less than 1.5°after calibration and fusion,the absolute value of the maximum error is less than 3°,and the calculated static angle error of the finger joint is less than 1°.The data glove system can detect the posture of the finger accurately,control the movement of the virtual dexterous hand in real time,and can also obtain the angle and angular velocity of the finger joint movement.

关 键 词：机器人遥操作 数据手套 惯性测量单元 Simulink 灵巧手 椭球拟合算法 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]