Vision based simultaneous localization and mapping using Sigma Point Kalman Filter

Simultaneous localization and mapping (SLAM) is one of the challenging issues in recent decades. In this paper solving vision based SLAM problem using Kalman filters family have been provided. It is focused on mobile robot equipped with stereo vision sensor which moves in an indoor environment. The mobile robot navigated among the landmarks which were detected by scale invariant feature transform (SIFT) method. The Extended Kalman Filter (EKF) approaches have been used to solve this SLAM problem. Then the role of sigma points in this filter to improve estimation accuracy of state in SLAM has been investigated. Finally the implementation results were presented to validate a better estimation of the state by Sigma Point Kalman Filter (SPKF) algorithm and its superiority over the EKF as a new method for solving the SLAM problem.