Observer and IMU-based detection and isolation of faults in position reference systems and gyrocompasses with dual redundancy in dynamic positioning

Faults in position reference systems and vessel sensors pose a serious hazard during dynamic positioning, and vessel drive-offs because of erroneous data may have dire consequences. The classification society standards require triple redundancy in heading and position reference systems, but common mode failures and operational limitations could put the safety promised by the redundancy at risk. In this paper we present a way to detect and isolate faults in the aforementioned systems, using nonlinear observers and inertial measurement units (IMUs). Employing IMU acceleration and gyro data in an integration filter, and comparing estimation error outputs from multiple observers in parallel proves to be a viable method for fault detection and isolation. This exploits the possibility of using IMUs as another, independent position reference system or sensor on board a vessel.