Range sensor-based localization of mobile robots in semi-structured environments

Abstract

There are various localization schemes for autonomous navigation of indoor mobile robots. However, most range sensor-based localization schemes only operate efficiently in environments where the geometric structure is fixed. In other words, localization performance is guaranteed only when the geometric structure of the actual environment matches with the geometric information given in the map. To resolve this limitation, this paper proposes a range sensor-based localization algorithm that can be efficiently used in semi-structured environments. By using both the natural and artificial landmarks simultaneously, a localization technique that is robust to environmental changes can be implemented. Experimental results show that localization performance in semi-structured environments can be improved using the proposed method.