UKF sensor fusion method based on principal component analysis
Jianye Yang, Dang Shu-wen, Fajiang He, Cheng Peng-zhan · Proceedings of the 3rd International Conference on Communication and Information Processing · 2017
In the process of mobile robot simultaneous localization and map building, to solve the problems, such as the information source of the laser radar navigation system being single and the assigned weight of multi-sensor fusion algorithm being unreasonable, a new UKF multi-sensor data fusion algorithm combined with principal component analysis (PCA) is proposed. In this PCA-UKF algorithm, the PCA based on multivariate statistical theory is used to distribute the weight deduced from the various sensors during navigation and calculate the state estimation after each measurement. Then, the estimated values which close to the real state are integrated into the observations. The experimental results show that the proposed algorithm can effectively improve the navigation accuracy and reliability. Furthermore, it performs better at fault tolerance and environment adaptability.