Modified Single Propagation Unscented Kalman Filter

Meng-de ZHANG, Kai-long LI, Bai-qing HU


When using single-point spread unscented Kalman filter, it is found that there may be large errors in estimating Jacobian matrix model with constant, resulting in poor estimation accuracy and even divergence. An improved single-point propagation unscented Kalman filter algorithm is proposed to solve this problem. By introducing the scaling factor, the influence of large distance between sampling points on the center point is reduced. Furthermore, the estimation accuracy and robustness of the single-point spread unscented Kalman filter algorithm are improved. In the simulation experiment, the ungm model is estimated. The performance of the single-point spread unscented Kalman filter, the unscented Kalman filter and the extended Kalman filter with different scaling factors are compared. A better scaling factor value is obtained. The improved algorithm is used to solve the measured data of indoor positioning, which proves the effectiveness of the algorithm.


Filter, Jacobian matrix, Sigma pointpropagation, Scaling factor


Full Text:



  • There are currently no refbacks.