Sensor fault detection and diagnosis for autonomous systems

Abstract

Autonomous systems are usually equipped with sensors to sense the surrounding environment. The sensor readings are interpreted into beliefs upon which the robot decides how to act. Unfortunately, sensors are susceptible to faults. These faults might lead to task failure. Detecting these faults and diagnosing a fault's origin is an important task that should be performed quickly online. While other methods require a high fidelity model that describes the behavior of each component, we present a method that uses a structural model to successfully detect and diagnose sensor faults online. We experiment our method with a laboratory robot Robotican1 and a flight simulator FlightGear. We show that our method outperforms previous methods in terms of fault detection and provides an accurate diagnosis.