LOCATION AND ORIENTATION TRACKING IN AUTONOMOUS SURFACE VEHICLES: REAL WORLD DATA VERSUS DIGITAL TWIN PREDICTIONS

The rapid advancement in autonomous technologies has driven the development and deployment of Digital Twins (DT) for various applications, including Autonomous Surface Vehicles (ASVs). This study focuses on creating and implementing a DT for an ASV to enhance environmental monitoring. The DT acts as a virtual replica of the ASV, simulating real-time data and predicting the vehicle's future state. The motivation for this project comes from the increasing need for precise and efficient environmental monitoring solutions while keeping track of the state of the ASV. The primary objective is to develop a DT for an ASV and evaluate its performance in predicting the vehicle's state in scenarios where sensor data is unavailable. The ASV and its DT were deployed in a controlled environment, with various performance metrics, including position and orientation, which are recorded and analyzed to assess the DT's predictive accuracy. When the DT received frequent updates, the mean absolute errors were relatively low: 1.411 meters for distance, 0.708 degrees for roll, 1.812 degrees for pitch, and 18.754 degrees for yaw. Extending the intervals between updates significantly affected prediction accuracy, with the MAE increased by 0.78% for roll, decreased by 21.42% for pitch, increased by 40.18% for yaw, and most significantly, the distance MAE was increased by 89.23%. The results of the study indicated that critical importance of advanced data fusion techniques and real-time processing capabilities from various sensors as well as control systems in real-time to maintain the reliability of the DT.