A Novel Slam Algorithm With Adaptive Kalman Filter

Xiangyuan Jiang,Tingting Li,Yunhua Yu

DOI: https://doi.org/10.1109/ICARM.2016.7606903

2016-01-01

Abstract:Simultaneous localization and map building(SLAM) concerns with filtering problem, which builds the map of the environment without robot's certainty position, meanwhile localize itself based on the map. Traditional SLAM algorithm addresses the problem by the Extended Kalman Filter which usually induces divergence because of inaccurate models. This paper proposes a novel SLAM algorithm which could overcome the problem of model mismatch, and modify the accuracy of localization and mapping further more.