Modeling of AUV Dynamics with Limited Resources: Efficient Online Learning Using Uncertainty

2025-04-06

AI Summary

To address high data redundancy and severe storage bottlenecks in online learning of autonomous underwater vehicle (AUV) dynamics under resource constraints, this paper proposes an uncertainty-driven online data selection framework. We explicitly quantify epistemic uncertainty via an ensemble of multilayer perceptrons and introduce three novel data filtering strategies—Threshold, Greedy, and Threshold-Greedy—first integrating uncertainty-aware selection into capacity-constrained online replay. The method jointly optimizes dynamic sampling, online gradient updates, and memory efficiency. Evaluated on real-world telemetry data from the Dagon AUV platform, the Threshold strategy significantly improves training stability and achieves the lowest cumulative test loss. Our analysis systematically uncovers the coupled influence of model complexity and memory budget on modeling accuracy, providing principled guidance for resource-aware adaptive modeling in embedded marine robotics.

Technology Category

• Machine Learning: Calibration & Uncertainty Quantification
• Intelligent Robots: State Estimation
• Reasoning under Uncertainty: Stochastic Optimization

Application Category

• Search and Retrieval-Augmented AI: Web learning to rank, online learning, and counterfactual learning for ranking
• Economics, Online Markets and Human Computation: Data quality aspects of human-annotated datasets
• Responsible Web: Machine-in-the-loop, human agency and autonomy

Abstract

Machine learning proves effective in constructing dynamics models from data, especially for underwater vehicles. Continuous refinement of these models using incoming data streams, however, often requires storage of an overwhelming amount of redundant data. This work investigates the use of uncertainty in the selection of data points to rehearse in online learning when storage capacity is constrained. The models are learned using an ensemble of multilayer perceptrons as they perform well at predicting epistemic uncertainty. We present three novel approaches: the Threshold method, which excludes samples with uncertainty below a specified threshold, the Greedy method, designed to maximize uncertainty among the stored points, and Threshold-Greedy, which combines the previous two approaches. The methods are assessed on data collected by an underwater vehicle Dagon. Comparison with baselines reveals that the Threshold exhibits enhanced stability throughout the learning process and also yields a model with the least cumulative testing loss. We also conducted detailed analyses on the impact of model parameters and storage size on the performance of the models, as well as a comparison of three different uncertainty estimation methods.

Problem

• Efficient online learning for AUV dynamics with limited storage
• Data selection using uncertainty to reduce redundant information
• Comparing novel uncertainty-based methods for model refinement

Innovation

• Uses uncertainty for data selection in online learning
• Employs ensemble of multilayer perceptrons for uncertainty prediction
• Introduces Threshold, Greedy, and Threshold-Greedy methods