改进的平方根容积模糊自适应卡尔曼滤波SLAM算法  

Improved Square Root Volume Fuzzy Adaptive Kalman Filter SLAM Algorithm

作　　者：李俊 舒志兵 王苏洲 

机构地区：[1]南京工业大学电气工程与控制科学学院,南京211816

出　　处：《组合机床与自动化加工技术》2017年第12期29-32,共4页Modular Machine Tool & Automatic Manufacturing Technique

摘　　要：针对移动机器人SLAM算法存在系统噪声对定位精度影响严重,特征点的增加导致轨迹偏移等现象。文章将迭代思想与时变渐消因子引入平方根容积卡尔曼滤波中,通过动态调节新息均值和协方差的方式,建立模糊自适应模型调整噪声权值,改善系统中存在的运动噪声和观测噪声。该算法相对于以往算法只能解决单一问题而言,具有更好的兼容性与鲁棒性。通过实验仿真结果可以看出,该算法相对于以往算法在X方向、Y方向和位姿偏移角的误差分别减小了21.59%、36.45%、32.97%。将此算法应用于实际中,具有良好的地图重建效果。SLAMalgorithm for mobile robot system is severely affected by the noise on the positioning accuracy,the phenomenon such as the increase of feature point lead to track migration. In this paper,the iterative thoughts and time-varying fading factor is introduced into the square root kalman filtering in the volume,and through dynamic adjust the way of the newinterest rates mean and covariance,and fuzzy adaptive adjust the weights of noise model,which improved the system noise and observation noise existing in the movement. The algorithm is compared with the previous algorithm can only solve a single problem,which has better compatibility and robustness. The simulation results showthat the proposed algorithm compared with previous algorithms in the X direction,Y direction error and position deviation angle decreases by 21. 59%,36. 45%,32. 97%. This algorithm is applied in practice,and has a good map reconstruction effect.

关 键 词：SLAM算法 迭代平方根容积 时变渐消因子 模糊自适应 

分 类 号：TH166[机械工程—机械制造及自动化] TG506[金属学及工艺—金属切削加工及机床]