作者：陈朋, 陈伟海, 于守谦

作者单位：北京航空航天大学自动化科学与电气工程学院, 北京 100191

关键词：状态方程; 量测方程; 信息融合; 无味卡尔曼滤波

摘要：

为提高室内轮式机器人定位精度,采用多传感器信息融合的机器人自主定位方法,根据室内轮式移动机器人的运动模型,建立了定位系统的状态方程;基于传感器的工作原理和数学模型,建立了各自的观测方程.鉴于二者的非线性,利用无味卡尔曼滤波算法对传感器信息进行融合.实验中开发了基于FPGA的主控平台,降低了数据处理系统的冗余度,提高了系统的稳定性.另外,设计了多传感器并行采集与快速处理的算法,提高了传感器信息融合的实时性和有效性.通过进行的机器人行走实验,结果表明该算法明显减小了定位误差,有效地提高了定位精度.

Research of localization for indoor mobile robot based on unscented Kalman filter

CHEN Peng, CHEN Wei-hai, YU Shou-qian

School of Automation Science and Electrical Engineering, Beihang University, Beijing 100191, China

Abstract: A self-localization method based on multi-sensor information fusion was proposed to improve the localization accuracy of the indoor wheeled mobile robot. In this paper, the state equation of the localization system was made based on the kinematics model of the wheeled mobile robot, and the measurement equations were built based on the working principle and the mathematical model of the sensors. An unscented Kalman filter was used to fuse the sensor information in connection with the nonlinearity of the two equations. A hardware platform with single field-programmable gate array(FPGA) as the master chip was developed to reduce the redundancy and improve the stability of data processing system. Besides, an algorithm of parallel acquisition and fast processing for multi-sensor was designed, which can improve the real-time and effectiveness of sensor fusion. Lastly, the robot walking experiment was carried out. Result proves that the method can reduce the localization errors markedly and improve the localization accuracy effectively.

Keywords: state equation; measurement equation; information fusion; unscented Kalman filter

2011, 37(5): 1-5  收稿日期: 2010-10-19;收到修改稿日期: 2011-1-14

基金项目: 国家自然科学基金项目(61075075); 国家863计划项目(2008AA04Z210); 北京市自然科学基金项目(3093021)

作者简介: 陈朋(1985-), 男, 河南商丘市人, 硕士研究生, 专业方向为智能机器人、嵌入式系统.

参考文献

[1] Woo J, Kim Y J, Lee J, et al. Localization of mobile robot using particle filter[C]//SICE-ICASE International Joint Conference. Busan, 2006: 3031-3034.
[2] 陈小宁, 黄玉清, 杨佳. 多传感器信息融合在移动机器人定位中的应用[J]. 传感器与微系统, 2008, 27(6): 110-113.
[3] 黄玉清, 张江梅, 张华. 模糊控制在移动机器人中的应用[J].西南科技大学学报, 2004, 19(1): 22-25.
[4] Drocout C, Delahoche L, Pegard C. Mobile robot localiza-tion based on an omnidirectional stereoscopic vision perception system[J].Robotics and Automation, 1999(2): 1322-1328.
[5] 蔡鹤皋, 金明河, 金峰. 卡尔曼滤波与多传感器数据融合技术[J]. 模式识别与人工智能, 2000, 13(2): 248-253.
[6] 宋骊平, 姬红兵. 多被动传感器UKF与EKF算法的应用与比较[J]. 系统工程与电子技术, 2009, 31(5): 1083-1086.
[7] Jetto L, Longhi S, Venturini G. Developmentand experim-
ental validation of an adaptive extended kalman filter for the localization of mobile robots[J]. IEEE Transactions on Robotics and Automation, 2006, 15(2): 219-229.
[8] Bao W, Zhang C W, Xiao B X, et al. Self-location of mobile robot based on binocular camera and unscented kalman filter[C]//International Conference on Automation and Logistics: Jinan, 2007: 277-281.
[9] Kwok N M, Liu D K, Dissanayake G.Evolutionary comp-
uting based mobile robot localization[J]. Engineering Applications of Arti.cial Intelligence, 2006, 19(8): 857-868.