摘要
针对外部干扰导致基于无迹卡尔曼滤波的同时定位与地图创建(UKF-SLAM)算法精度降低甚至发散的问题,提出一种改进的UKF-SLAM算法.算法在系统状态更新之前引入外部干扰检测和状态方差膨胀,在干扰发生的周期内快速做出响应,提高了系统状态估计的精度.临近观测的对比检测不受累计误差的影响,提高了检测的精度,不同状态方差膨胀方式能够响应不同类型的外部干扰.仿真实验结果表明,在存在外部干扰的环境中,改进UKF-SLAM算法估计精度高于SMCI-SLAM和UKF-SLAM算法.
Outlier disturbances decreased accuracy of unscented Kalman filter for simultaneous localization and mapping(UKF-SLAM),and they even led UKF-SLAM to divergence.An improved UKFSLAM algorithm was proposed to solve this problem.This algorithm introduced outlier disturbance detection and covariance inflation to UKF-SLAM before the process of system state updating.It can response outlier disturbances during the same period,and improve the robustness.The comparison of time-adjacent observations improved the detection accuracy without cumulative errors.When outlier disturbances were detected,different expansion methods of system covariance matrix were adapted to different outlier disturbances.Simulation results show that the proposed algorithm gets more accurate estimation than SMCI-SLAM and UKF-SLAM with outlier disturbances.
作者
吕太之
周武
赵春霞
LU Tai-zhi;ZHOU Wu;ZHAO Chun-xia(School of Information Technology,Jiangsu Maritime Institute,Nanjing 211170,China;College of Computer Science and Technology,Nanjing University of Science and Technology,Nanjing 210094,China;School of Engineering,Zhejiang Normal University,Jinhua 321004,China)
出处
《中北大学学报（自然科学版）》
CAS
2018年第6期717-725,751,共10页
Journal of North University of China(Natural Science Edition)
基金
国家自然科学基金青年科学基金项目(51405450)
江苏省高校"青蓝工程"优秀青年骨干教师
江苏省高校优秀中青年教师和校长境外研修项目
江苏海事学院千帆团队建设项目
关键词
移动机器人
同时定位与地图创建
无迹卡尔曼滤波
外部干扰
mobile robot
simultaneous localization and mapping
unscented Kalman filter
outlier disturbance
