Leveraging digital twins for fault diagnosis in autonomous ships

Autonomous ships are vulnerable to a spectrum of potential faults, including propeller damage, shaft-line failures, and electrical supply disruptions. Swift fault detection is critical to minimize the potentially catastrophic consequences of these issues. In this paper, we introduce an innovative digital twin-based method for diagnosing faults in autonomous ships. Our digital twin system collects data from the ship's sensor array, conducts comprehensive analysis, computes fault parameters, and provides real-time visualizations of the findings. To achieve this, we utilize the ship's dynamical model and employ a novel adaptive extended Kalman filter (AEKF) algorithm to estimate the severity of these faults. We evaluate the effectiveness of this approach through numerical simulations and practical implementation in an autonomous surface vehicle called the Otter, developed by Maritime Robotics. The experimental results, which consider both normal and faulty ship propulsion systems, underscore the significant potential of this novel approach to improve ship condition monitoring during operational activities.