Iterative filtering based on Kazakov linearization

摘要

The research on nonlinear filtering algorithm is an active orientation. In this paper, a novel iterative filtering algorithm based on Kazakov linearization is proposed, and the derivation of the algorithm based the Gauss-Newton method is addressed in detail. This algorithm is applied to computer simulation of the SLAM problem in the robot location field and simulation result has been compared with UKF algorithm. The simulation shows that Kazakov based nonlinear filter perform better than UKF algorithm.