基于改进容积卡尔曼滤波器的超宽带/惯性导航系统室内融合定位技术

Indoor fusion positioning technology of ultra wide band/inertial navigation system based on improved cubature Kalman filter

在室内非视距环境中,单一超宽带(ultra wide band,UWB)技术会产生稳定性差和定位精度低的问题,以至于不能达到室内定位的要求。将UWB定位技术和惯性导航系统(inertial navigation system,INS)技术相结合,能够有效抑制非视距误差对定位结果的影响,因此本文提出基于改进的容积卡尔曼滤波器(cubature Kalman filter,CKF)的UWB/INS融合室内定位方法。首先,通过广义似然比检测算法对行人的静止态和运动态进行检测;之后对于静止态通过零速修正算法对速度进行修正,零积分航向角速率修正算法对航向角进行修正;然后,通过模糊综合评判的非视距识别算法对非视距信号进行识别,通过基于半定规划的非视距修正算法对识别出的信号进行修正;最后通过改进的容积卡尔曼滤波器将UWB定位数据和INS定位数据进行融合,得出组合定位结果。实验结果显示,与标准容积卡尔曼滤波算法相比,改进的容积卡尔曼滤波算法的定位结果均方根误差降低了2.32 cm(18.82%),拥有更好的定位精度和鲁棒性。

In an indoor non-line-of-sight environment,a single ultra wide band(UWB)technique can lead to poor stability and low positioning accuracy,so that it cannot meet the requirements of indoor positioning.The combination of UWB positioning technology and inertial navigation system(INS)technology can effectively suppress the influence of non-line-of-sight error on the positioning results.Therefore,a UWB/INS fused indoor positioning method based on improved cubature Kalman filter(CKF)is proposed.Firstly,the general likelihood ratio detection algorithm is used to detect the stationary and moving states of pedestrians.Then,for the stationary state,the speed is corrected using the zero velocity update algorithm,and the heading angle is corrected using the zero integral heading rate correction algorithm.Then,the non-line-of-sight signals are recognized by the non-line-of-sight recognition algorithm based on fuzzy comprehensive evaluation,and next,the identified signals are corrected by the non-line-of-sight correction algorithm based on semi-definite programming.Finally,the UWB positioning data and INS positioning data are fused by the improved cubature Kalman filter to obtain the combined positioning result.The experimental results show that compared with the standard cubature Kalman filter algorithm,the root-mean-square deviation of the improved cubature Kalman filter filter algorithm is reduced by 2.32 cm(18.82%),having better positioning accuracy and robustness.