Sensors Faults Detection and Isolation using EKF-SLAM for a Mobile Robot

The aim of this paper is to accomplish the mobile robots full navigation by developing tools that are able to give a solution in faulty situations. It treats Simultaneous Localization and Mapping (SLAM) problem based on the Extended Kalman Filter (EKF) using proprioceptive and exteroceptive sensors affected by hardware faults. In this work, sensor Fault Detection and Isolation (FDI) approach using SLAM fusion is proposed which is based on a traditional method (duplication/comparison) when an absolute localization sensor is available. In order to gather accurate information about the FDI approach, mobile robot is equipped with a variety of sensors such as encoders, gyroscope, laser rangefinder and indoor GPS. The proposed approach was validated under realistic conditions using experimental data and good performances were obtained.