Fault Detection in Inertial Measurement Unit and Global Navigation Satellite System of an Unmanned surface vehicle

This paper presents a method which detects faults in the measurement required for navigation of an unmanned surface vehicle. The measurement includes acceleration and angular rate by an inertial measurement unit (IMU), and the location by global navigation satellite system (GNSS). The method uses Kalman filter to estimate the state variables which represent the amount of faulty measurements. The method is tested by the experiment in which a boat equipped by commercial IMU and GNSS navigates in a sea trial. The result verifies that faults in acceleration and angular rate measurements are detected robustly and the GNSS position fault is detected when there is considerable location error.