Sensor failure detection in autonomous mobile robot application

Presented paper deals with sensor failure detection and identification (FDI) in navigation task of autonomous mobile robot. Mobile robots use ultrasonic range finders to solve a collision avoidance task, which is crucial for successful navigation. When the sensory readings are invalid the robot should be able to detect such a failure and interrupt its actual movement. To solve this task, the method based on first-order Markov chain parameters estimation was developed and tested in both the simulation and the real world environment. Practical experiments were performed on autonomous mobile robot Advee.