UWB精确定位自适应无迹卡尔曼滤波算法

An Adaptive Unscented Kalman Filter Algorithm for UWB Precise Positioning

针对超宽带定位(UWB)过程中的非视距(NLOS)误差和非线性滤波问题,提出一种基于峭度识别的UWB精确定位自适应无迹卡尔曼滤波算法。通过分析基于到达时间定位原理,建立了NLOS误差模型。使用峭度作为衡量UWB信道环境的参数,设定阈值鉴别NLOS误差。在无迹卡尔曼滤波过程中引入自适应因子,利用残差向量不断修正增益矩阵,从而减小误差。试验证明,视距环境下,此算法定位精度可达厘米级;非视距环境下,定位精度达到亚分米级。

Aiming at the NLOS error and nonlinear filtering in UWB positioning,an adaptive unscented Kalman filter algorithm based on kurtosis recognition is proposed.NLOS error model is established by analyzing the principle of arrival time location.Kurtosis is used as a parameter to measure UWB channel environment,and threshold is set to identify NLOS error.In the process of unscented Kalman filter,an adaptive factor is introduced,and the gain matrix is modified by the residual vector to reduce the error.Experimental results show that the positioning accuracy of this algorithm can reach centimeter level in LOS environment and sub decimeter level in NLOS environment.