DVL Calibration using Data-driven Methods

TL;DR

This work tackles DVL calibration for AUVs in GNSS-denied underwater environments by introducing an end-to-end data-driven calibration framework. A dual-CNN architecture with a complementary 1D head processes DVL and GNSS-RTK velocity data to estimate axiswise DVL error terms under four models, including per-axis scale and bias. Trained on large-scale simulated data, the approach reduces calibration time and, for low-end DVLs, improves RMSE relative to a model-based baseline, with EM4 (per-axis bias) yielding the most notable gains. The results highlight the potential of data-driven calibration to simplify and accelerate DVL calibration in practical underwater navigation, while pointing to the need for real-world validation.

Abstract

Autonomous underwater vehicles (AUVs) are used in a wide range of underwater applications, ranging from seafloor mapping to industrial operations. While underwater, the AUV navigation solution commonly relies on the fusion between inertial sensors and Doppler velocity logs (DVL). To achieve accurate DVL measurements a calibration procedure should be conducted before the mission begins. Model-based calibration approaches include filtering approaches utilizing global navigation satellite system signals. In this paper, we propose an end-to-end deep-learning framework for the calibration procedure. Using stimulative data, we show that our proposed approach outperforms model-based approaches by 35% in accuracy and 80% in the required calibration time.

Figures (3)

• Figure 1: Four DVL error models examined in this work.
• Figure 2: Block diagram illustration of our architecture. The DVL and GNSS-RTK velocity measurements are fed and processed simultaneously by the two convolution heads, the output is concatenated and fed through the FC head which outputs the estimated DVL error term.
• Figure 3: A simulation pipeline for generating DVL and GNSS measurements.