Uncertainty Estimation for Deep Reconstruction in Actuatic Disaster Scenarios with Autonomous Vehicles

TL;DR

This paper evaluates various uncertainty estimation methods for environmental scalar field reconstruction in aquatic monitoring, highlighting Evidential Deep Learning as the most effective and efficient approach for autonomous vehicles.

Contribution

It provides a comprehensive comparison of uncertainty quantification techniques, demonstrating Evidential Deep Learning's superiority in accuracy, calibration, and computational efficiency.

Findings

Evidential Deep Learning outperforms other methods in reconstruction accuracy.

Gaussian Processes are limited by stationary kernel assumptions and scalability issues.

Evidential Deep Learning offers the best trade-off between accuracy and inference cost.

Abstract

Accurate reconstruction of environmental scalar fields from sparse onboard observations is essential for autonomous vehicles engaged in aquatic monitoring. Beyond point estimates, principled uncertainty quantification is critical for active sensing strategies such as Informative Path Planning, where epistemic uncertainty drives data collection decisions. This paper compares Gaussian Processes, Monte Carlo Dropout, Deep Ensembles, and Evidential Deep Learning for simultaneous scalar field reconstruction and uncertainty decomposition under three perceptual models representative of real sensor modalities. Results show that Evidential Deep Learning achieves the best reconstruction accuracy and uncertainty calibration across all sensor configurations at the lowest inference cost, while Gaussian Processes are fundamentally limited by their stationary kernel assumption and become intractable as observation density grows. These findings support Evidential Deep Learning as the preferred method for uncertainty-aware field reconstruction in real-time autonomous vehicle deployments.