A Hybrid Data Association Approach for SLAM in Dynamic Environments

Abstract

Data association is critical for Simultaneous Localization and Mapping (SLAM). In a real environment, dynamic obstacles will lead to false data associations which compromise SLAM results. This paper presents a simple and effective data association method for SLAM in dynamic environments. A hybrid approach of data association based on local maps by combining ICNN and JCBB algorithms is used initially. Secondly, we set a judging condition of outlier features in association assumptions and then the static and dynamic features are detected according to spatial and temporal difference. Finally, association assumptions are updated by filtering out the dynamic features. Simulations and experimental results show that this method is feasible.