两轮自平衡机器人惯性传感器滤波问题的研究  

Research on Filtering Problem in Inertial Sensors for a Two-Wheeled Self-Balanced Robot

作　　者：郜园园[1] 阮晓钢[1] 宋洪军[1] 陈静[1] 

机构地区：[1]北京工业大学电子信息与控制工程学院,北京100124

出　　处：《传感技术学报》2010年第5期696-700,共5页Chinese Journal of Sensors and Actuators

基　　金：国家自然科学基金资助项目资助(60774077);国家"863计划"资助项目资助(2007AA04Z226)

摘　　要：针对惯性传感器在两轮机器人姿态检测中存在随机漂移误差的问题,基于卡尔曼滤波实现对倾角仪与陀螺仪的信息融合,设计了简单而实用的滤波算法,对传感器的误差进行补偿后得到机器人姿态信号的最优估计,从而将其应用于两轮自平衡机器人系统。实验结果表明,采用卡尔曼信息融合的方法,来得到机器人姿态信息最优估计是有效可行的,并且有利于机器人完成自平衡的控制。Aiming at the random drift error from inertial sensors of a two-wheeled self-balanced robot attitude measuring,a simple and practical filtering algorithm based on Kalman filter which was implemented to information fusion for inclinometer and gyroscope was proposed,thus realizing optimal estimation for the robot gesture signal after sensors error compensation.The experimental results showed that the method based on Kalman information fusion to obtain the optimal estimation was effective and feasible.It is also beneficial to complete the robot self-balancing control.

关 键 词：姿态检测 信息融合 卡尔曼滤波 惯性传感器 

分 类 号：TP212.9[自动化与计算机技术—检测技术与自动化装置]