Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Deep Learning Powered Estimate of The Extrinsic Parameters on Unmanned Surface Vehicles

Yi Shen, Hao Liu, Chang Zhou, Wentao Wang, Zijun Gao, Qi Wang

TL;DR

This work tackles real-time extrinsic-parameter calibration for USVs in dynamic marine environments by predicting metacenter shifts from Euler angles with a Time-Sequence GRNN. The approach compares FC, RBF, GRNN, and a time-sequence extension, showing the Time-Sequence GRNN achieving the lowest MSE in testing ($L_{MSE}$ values of $6.65$) and a $2.58\ \text{cm}$ metacenter error at a USV length of $9.5\ \text{m}$, trained on Unity3D simulations. The method leverages temporal context to model dynamic stability effects, addressing the computational bottleneck of FEM-based FEM mappings with a neural surrogate. The results indicate strong real-time calibration potential in dynamic conditions, though real-world validation remains a key next step to confirm robustness outside simulation.

Abstract

Unmanned Surface Vehicles (USVs) are pivotal in marine exploration, but their sensors' accuracy is compromised by the dynamic marine environment. Traditional calibration methods fall short in these conditions. This paper introduces a deep learning architecture that predicts changes in the USV's dynamic metacenter and refines sensors' extrinsic parameters in real time using a Time-Sequence General Regression Neural Network (GRNN) with Euler angles as input. Simulation data from Unity3D ensures robust training and testing. Experimental results show that the Time-Sequence GRNN achieves the lowest mean squared error (MSE) loss, outperforming traditional neural networks. This method significantly enhances sensor calibration for USVs, promising improved data accuracy in challenging maritime conditions. Future work will refine the network and validate results with real-world data.

Deep Learning Powered Estimate of The Extrinsic Parameters on Unmanned Surface Vehicles

TL;DR

This work tackles real-time extrinsic-parameter calibration for USVs in dynamic marine environments by predicting metacenter shifts from Euler angles with a Time-Sequence GRNN. The approach compares FC, RBF, GRNN, and a time-sequence extension, showing the Time-Sequence GRNN achieving the lowest MSE in testing ( values of ) and a metacenter error at a USV length of , trained on Unity3D simulations. The method leverages temporal context to model dynamic stability effects, addressing the computational bottleneck of FEM-based FEM mappings with a neural surrogate. The results indicate strong real-time calibration potential in dynamic conditions, though real-world validation remains a key next step to confirm robustness outside simulation.

Abstract

Unmanned Surface Vehicles (USVs) are pivotal in marine exploration, but their sensors' accuracy is compromised by the dynamic marine environment. Traditional calibration methods fall short in these conditions. This paper introduces a deep learning architecture that predicts changes in the USV's dynamic metacenter and refines sensors' extrinsic parameters in real time using a Time-Sequence General Regression Neural Network (GRNN) with Euler angles as input. Simulation data from Unity3D ensures robust training and testing. Experimental results show that the Time-Sequence GRNN achieves the lowest mean squared error (MSE) loss, outperforming traditional neural networks. This method significantly enhances sensor calibration for USVs, promising improved data accuracy in challenging maritime conditions. Future work will refine the network and validate results with real-world data.
Paper Structure (25 sections, 2 equations, 8 figures, 1 table)

This paper contains 25 sections, 2 equations, 8 figures, 1 table.

Table of Contents

  1. Introduction
  2. Related Work
  3. Metacenter
  4. Extrinsic Parameters Estimate on USVs
  5. Method and Implementation
  6. Radial basis function network
  7. Gaussian kernal function
  8. General regression neural network
  9. Time sequence information
  10. Loss function
  11. Experiments and results
  12. Origin of data
  13. Data collection and process
  14. Results
  15. Fully connected network
  16. ...and 10 more sections

Figures (8)

  • Figure 1: Ship stability condition.vasilescu2020influence
  • Figure 2: The structure of the RBF.
  • Figure 3: The structure of the GRNN.
  • Figure 4: The structure of the time sequence information network.
  • Figure 5: MSE loss of the fully connected network
  • ...and 3 more figures