Improved Unscented Kalman Filtering Algorithm Applied to On-vehicle Tracking System
Yun Ruo Chen, Yi Liu · 2021
In this paper, we study the problem of tracking aerial targets in a modified spherical coordinate system using an on-vehicle tracking system using infrared sensors. In this challenging scenario, to solve the problems of poor calculation stability and large amount of calculation of the traditional unscented kalman filter algorithm, this article introduces and adopts an improved unscented kalman filter algorithm. In Sigma sampling, the error variance matrix uses singular value decomposition (SVD) to select the simplex with the smallest skewness. Compared with the estimation algorithm based on the standard unscented kalman filter, the new algorithm effectively reduces the state estimation error and improves the tracking stability and tracking accuracy.