Enhancing seamless INS/UWB integrated localization using LS-SVM assisted predictive adaptive Kalman filter

This paper presents a least squares-support vector machines (LS-SVM) assisted predictive Kalman filter (PKF). In this model, the PKF can fuse the ultra wide band (UWB)- and inertial navigation system (INS)-based measurement, and compensate the missing data. Moreover, the LS-SVM is used to build the mapping between theoretical values and observation vectors, which can improve the accuracy. The experimental study indicates the effectiveness of the proposed LS-SVM assist PKF.