Iterated Unscented SLAM algorithm for navigation of an autonomous mobile robot

Unscented Kalman Filter (UKF) is one of the most frequently used nonlinear estimators from the view point of estimation accuracy and easy implementation to solve the SLAM problem that is often referred to as Unscented SLAM algorithm. This paper investigates the possibility of reduction of estimation error due to statistical linearization of nonlinear measurement model in Unscented SLAM (USLAM) algorithm. We take advantage of an iteration mechanism in update equations of Unscented SLAM in order to reduce the statistical error propagation existing in this algorithm. In this paper, Simulation results have shown better performance of the Iterated Unscented SLAM (IUSLAM) algorithm. Finally, simulation results are consistently validated by real-world experiments based on collected data from a mobile robot in our laboratory.