摘要
针对移动机器人位置的精确估计问题,提出一种附有约束的无色卡尔曼滤波(CUKF)算法。由于无色卡尔曼滤波(UKF)在处理非线性问题时,无需计算Jacobian矩阵或Hessian矩阵,从而有效地减小了线性化对非线性系统误差的影响。CUKF算法很好地利用了UKF的非线性滤波特点,在其基础上增加某种约束。将地理信息系统(GIS)环境下的地图数据库中的道路方向信息作为约束条件,通过引入拉格朗日函数解决具有约束的差分全球定位系统/航位推算(DGPS/DR)组合导航系统的非线性最优估计。仿真实验结果表明:CUKF比UKF能够更有效地提高定位精度。
The algorithm about a constrained unscented Kalman filter （CUKF）is proposed to esti mate the localization of the mobile robot. The unscented Kalman filter （UKF） doesn＇t compute the Jacobian matrix nor the Hessian matrix when it processes the nonlinear questions and it can decrease effectively the error of nonlinear system caused by the linearization. The algorithm of CUKF takes full advantage of characteristics of the UKF and some restrictions are increased. The road direction of the digital map database in geographic information system （GIS） environment is regarded as the state constrained condition. The nonlinear optimization estimation about DGPS/DR （ difference glob- al positioning system/dead reckoning） integrated navigation system can be resolved through the Lagrangian function. Simulation results show that the CUKF can more effectively improve the position precision than the UKF.
出处
《南京理工大学学报》
EI
CAS
CSCD
北大核心
2009年第1期37-41,共5页
Journal of Nanjing University of Science and Technology
基金
国家自然科学基金(60705020)
关键词
无色卡尔曼滤波
移动机器人
组合导航
差分全球定位系统
航位推算
unscented Kalman filter
mobile robots
integrated navigation
difference global positioning system
dead reckoning
