移动机器人的多传感器信息融合  

Multi-sensor information fusion for mobile robots

作　　者：李永强 唐旭东[1] 李兆凯[1] 周云虎 LI Yongqiang;TANG Xudong;LI Zhaokai;Zhou Yunhu(Beijing Institute of Mechanical Equipment,Beijing 100854,China;Department Mechatronics Engineering,Harbin Institute of Technology,Harbin 150001,China)

机构地区：[1]北京机械设备研究所,北京100854 [2]哈尔滨工业大学机械电子工程系,黑龙江哈尔滨150001

出　　处：《西北工业大学学报》2021年第S01期59-65,共7页Journal of Northwestern Polytechnical University

摘　　要：为了提高移动机器人的感知能力,使机器人具有更高的自主性,在移动机器人需要感知姿态的部位装配MEMS传感器,包括三轴加速度计和陀螺仪。可以通过对其读数进行数据融合,得到机器人的位姿信息。通过互补滤波算法、扩展卡尔曼滤波算法进行传感器信息融合,采用四元数法得到机器人关节的位姿,为机器人对自身状态估计以及在复杂环境下做出相应的决策提供依据。以Vicon三维运动系统中记录的位姿变化为标准,对2种算法的准确性进行对比。In order to improve the perception ability of the mobile robot and make the robot have higher autonomy,MEMS sensors,including three-axis accelerometer and gyroscope,are installed in the position of the mobile robot that needs to perceive the attitude.The position and pose information of the robot can be obtained by data fusion of its readings.In this paper,the complementary filtering algorithm and the extended Kalman filtering algorithm are used for sensor information fusion,and the quaternion method is used to obtain the position and pose of the robot joint,which provides the basis for the robot to estimate its own state and make the corresponding decision in the complex environment.Finally,the accuracy of the two algorithms is compared with the changes of position and pose recorded in VICON 3D motion system.

关 键 词：互补滤波算法 扩展卡尔曼滤波算法 多传感器信息融合 

分 类 号：TP242.6[自动化与计算机技术—检测技术与自动化装置]