Digital Twin-based Out-of-Distribution Detection in Autonomous Vessels

TL;DR

ODDIT addresses real-time out-of-distribution detection in autonomous vessels by building a data-driven digital twin composed of a recurrent neural network model (DTM) to forecast future AV states and a deep autoencoder (DTC) to judge whether predicted states deviate from normal behavior. The framework uses an OOD threshold $T_{OOD}$, defined as $T_{OOD}=\mu(RE)+3\sigma(RE)$, to decide IND vs OOD based on reconstruction errors. Evaluated on five vessel models across waypoint and zigzag maneuvers with sensor/actuator noise and ocean-current disturbances, ODDIT achieves AUROC and TNR@TPR95 up to 99%, outperforming RMSE and Euclidean baselines in most scenarios, with some vessel- and disturbance-specific variability. The results demonstrate strong real-time detection capabilities and reveal the importance of vessel-specific tuning and robust training data, while outlining directions for adaptive thresholds, uncertainty quantification, and industrial case studies to enhance practical deployment.

Abstract

An autonomous vessel (AV) is a complex cyber-physical system (CPS) with software enabling many key functionalities, e.g., navigation software enables an AV to autonomously or semi-autonomously follow a path to its destination. Digital twins of such AVs enable advanced functionalities such as running what-if scenarios, performing predictive maintenance, and enabling fault diagnosis. Due to technological improvements, real-time analyses using continuous data from vessels' real-time operations have become increasingly possible. However, the literature has little explored developing advanced analyses in real-time data in AVs with digital twins built with machine learning techniques. To this end, we present a novel digital twin-based approach (ODDIT) to detect future out-of-distribution (OOD) states of an AV before reaching them, enabling proactive intervention. Such states may indicate anomalies requiring attention (e.g., manual correction by the ship master) and assist testers in scenario-centered testing. The digital twin consists of two machine-learning models predicting future vessel states and whether the predicted state will be OOD. We evaluated ODDIT with five vessels across waypoint and zigzag maneuvering under simulated conditions, including sensor and actuator noise and environmental disturbances i.e., ocean current. ODDIT achieved high accuracy in detecting OOD states, with AUROC and TNR@TPR95 scores reaching 99\% across multiple vessels.

Figures (12)

• Figure 1: A conceptual model for digital twins for cyber-physical systems.
• Figure 2: Overview of ODDIT, highlighting the DT and the corresponding physical system.
• Figure 3: Examples of waypoint navigation: (a) 3DoF vessel navigating in a 2D path, (b) 6DoF vessel navigating in a 3D path, and (c) OOD behavior due to noise in position sensors.
• Figure 4: Zigzag with normal 20° rudder angle.
• Figure 5: Zigzag with high 40° rudder angle.