notesum.ai

Published at November 22

Bayesian dynamic mode decomposition for real-time ship motion digital twinning

stat.AP

math.DS

Released Date: November 22, 2024

Authors: Giorgio Palma¹, Andrea Serani¹, Kevin McTaggart², Shawn Aram³, David W. Wundrow³, David Drazen, Matteo Diez¹

Aff.: ¹National Research Council-Institute of Marine Engineering, Rome, Italy; ²Defence Research and Development Canada-Atlantic Research Center, Dartmouth, Nova Scotia, Canada; ³Naval Surface Warfare Center-Carderock Division, West Bethesda, Maryland, U.S.A.

Arxiv: http://arxiv.org/abs/2411.14839v1

NRMSE			$l_{te}=\hat{T}$	$l_{te}=2\hat{T}$	$l_{te}=5\hat{T}$
CFDShip-Iowa	Best deterministic	avg	0.3329	0.4697	0.7417
		(std)	(0.1684)	(0.2101)	(0.2950)
	Bayesian	avg	0.2736	0.4061	0.6626
		(std)	(0.1222)	(0.1730)	(0.2628)
TEMPEST	Best deterministic	avg	0.4247	0.5284	0.7940
		(std)	(0.2422)	(0.2312)	(0.2882)
	Bayesian	avg	0.3744	0.5029	0.7653
		(std)	(0.1900)	(0.2237)	(0.2878)
ShipMo3D	Best deterministic	avg	0.4683	0.6161	0.9223
		(std)	(0.4018)	(0.3958)	(0.5283)
	Bayesian	avg	0.4066	0.5398	0.8045
		(std)	(0.3443)	(0.3316)	(0.3365)