U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis

This study introduces a Multiscale Dual-Attention U-Net (TS-MSDA U-Net) model for long-term time series synthesis. By integrating multiscale temporal feature extraction and dual-attention mechanisms into the U-Net backbone, the model captures complex temporal dependencies more effectively. The model...

Full description

Saved in:
Bibliographic Details
Main Author: Shyr-Long Jeng
Format: Article
Language:English
Published: MDPI AG 2025-06-01
Series:Sensors
Subjects:
Online Access:https://www.mdpi.com/1424-8220/25/13/4073
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1849701836138741760
author Shyr-Long Jeng
author_facet Shyr-Long Jeng
author_sort Shyr-Long Jeng
collection DOAJ
description This study introduces a Multiscale Dual-Attention U-Net (TS-MSDA U-Net) model for long-term time series synthesis. By integrating multiscale temporal feature extraction and dual-attention mechanisms into the U-Net backbone, the model captures complex temporal dependencies more effectively. The model was evaluated in two distinct applications. In the first, using multivariate datasets from 70 real-world electric vehicle (EV) trips, TS-MSDA U-Net achieved a mean absolute error below 1% across key parameters, including battery state of charge, voltage, acceleration, and torque—representing a two-fold improvement over the baseline TS-p2pGAN. While dual-attention modules provided only modest gains over the basic U-Net, the multiscale design enhanced overall performance. In the second application, the model was used to reconstruct high-resolution signals from low-speed analog-to-digital converter data in a prototype resonant CLLC half-bridge converter. TS-MSDA U-Net successfully learned nonlinear mappings and improved signal resolution by a factor of 36, outperforming the basic U-Net, which failed to recover essential waveform details. These results underscore the effectiveness of transformer-inspired U-Net architectures for high-fidelity multivariate time series modeling in both EV analytics and power electronics.
format Article
id doaj-art-85a8cc9f1d344c7ea168fce8e23293f1
institution DOAJ
issn 1424-8220
language English
publishDate 2025-06-01
publisher MDPI AG
record_format Article
series Sensors
spelling doaj-art-85a8cc9f1d344c7ea168fce8e23293f12025-08-20T03:17:51ZengMDPI AGSensors1424-82202025-06-012513407310.3390/s25134073U-Net Inspired Transformer Architecture for Multivariate Time Series SynthesisShyr-Long Jeng0Department of Mechanical Engineering, Lunghwa University of Science and Technology, Taoyuan City 333326, TaiwanThis study introduces a Multiscale Dual-Attention U-Net (TS-MSDA U-Net) model for long-term time series synthesis. By integrating multiscale temporal feature extraction and dual-attention mechanisms into the U-Net backbone, the model captures complex temporal dependencies more effectively. The model was evaluated in two distinct applications. In the first, using multivariate datasets from 70 real-world electric vehicle (EV) trips, TS-MSDA U-Net achieved a mean absolute error below 1% across key parameters, including battery state of charge, voltage, acceleration, and torque—representing a two-fold improvement over the baseline TS-p2pGAN. While dual-attention modules provided only modest gains over the basic U-Net, the multiscale design enhanced overall performance. In the second application, the model was used to reconstruct high-resolution signals from low-speed analog-to-digital converter data in a prototype resonant CLLC half-bridge converter. TS-MSDA U-Net successfully learned nonlinear mappings and improved signal resolution by a factor of 36, outperforming the basic U-Net, which failed to recover essential waveform details. These results underscore the effectiveness of transformer-inspired U-Net architectures for high-fidelity multivariate time series modeling in both EV analytics and power electronics.https://www.mdpi.com/1424-8220/25/13/4073attentionCLLC converterhalf-bridgetime series synthesiselectric vehicle
spellingShingle Shyr-Long Jeng
U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
Sensors
attention
CLLC converter
half-bridge
time series synthesis
electric vehicle
title U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
title_full U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
title_fullStr U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
title_full_unstemmed U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
title_short U-Net Inspired Transformer Architecture for Multivariate Time Series Synthesis
title_sort u net inspired transformer architecture for multivariate time series synthesis
topic attention
CLLC converter
half-bridge
time series synthesis
electric vehicle
url https://www.mdpi.com/1424-8220/25/13/4073
work_keys_str_mv AT shyrlongjeng unetinspiredtransformerarchitectureformultivariatetimeseriessynthesis