Skip to content
DraconiticResearch
Search Sources API
203,234 studies31 sourcesUpdated just now
Peer-reviewedCC BYPLOS One

UniHAR: A unified hierarchical transformer with multi-scale temporal modeling for wearable human activity recognition

  • Memoona1
  • Muhammad Haseeb Hashir1
  • Sung Won Kim2
  1. 1Department of Information and Communication Engineering, Yeungnam University, Gyeongsan, Gyeongbuk, Republic of Korea.
  2. 2School of Computer Science and Engineering, Yeungnam University, Gyeongsan, Gyeongbuk, Republic of Korea.
Read full textPDFDOICite

Abstract

UniHAR is a lightweight hierarchical multi-scale transformer based on wearable sensor data that has to be able to model time well; be able to recognize human activity across different datasets; and do so in real-time. There are several problems with what currently exists. UniHAR uses four major components: universal channel harmonization; adaptive multi-scale patch embedding; scale-adaptive Fourier cross-attention; factorized temporal-channel attention. The components allow for the ability to use various sensors and enables the low-complexity multi-scale fusion of features. UniHAR achieved an accuracy of 96.35%; 94.45%; 96.44%; and 99.23% on UCI-HAR, Opportunity, Daphnet, and PAMAP2 datasets, while being very lightweight and having high-speed inference capabilities; therefore, UniHAR is suitable for real-time healthcare applications.

Related

  1. Hierarchical Temporal Transformer for Cancer Grade Prediction and Cross-Cancer Transfer Learning from Pathology Reports
    n. Brimo, R. Anand, A. Sabath +2
    bioRxivPreprintCC BYFull textBioinformatics & computational biologybioRxiv
  2. Hierarchical temporal transformer for cancer grade prediction and cross cancer transfer learning from pathology reports
    N. Brimo, R. Anand, H. Harb +1
    bioRxivPreprintCC BYFull textBioinformatics & computational biologybioRxiv
  3. A context-aware temporal modeling through unified multi-scale temporal encoding and hierarchical sequence learning for single-channel EEG sleep staging
    Amir Ali Vakili, Salar Jahanshiri, Armin Salimi-Badr
    PLOS OnePeer-reviewedCC BY 4.0Full textNeurosciencePhysiologyPLOS
  4. Ourotide: decoding the hierarchical peptide recognition for generative design
    Y. Shen, J. Zhang, Z. Wu +7
    bioRxivPreprintAll rights reservedFull textBioinformatics & computational biologybioRxiv
  5. Few-shot recognition of lane-changing intention in connected vehicle environment: A meta-transformer approach
    Xiaoqian Liu, Wenxin Jiang, Yining Ren
    PLOS OnePeer-reviewedCC BYFull textNeuroscienceGeneral mathematicsPLOSEurope PMCPMC

Cite

BibTeXRIS
BibTeX
@article{memoona2026unihar,
  title = {{UniHAR: A unified hierarchical transformer with multi-scale temporal modeling for wearable human activity recognition}},
  author = {Memoona and Muhammad Haseeb Hashir and Sung Won Kim},
  journal = {PLOS One},
  year = {2026},
  month = oct,
  doi = {10.1371/journal.pone.0359774},
  url = {https://doi.org/10.1371/journal.pone.0359774},
  note = {CC BY},
}
RIS
TY  - JOUR
TI  - UniHAR: A unified hierarchical transformer with multi-scale temporal modeling for wearable human activity recognition
AU  - Memoona
AU  - Muhammad Haseeb Hashir
AU  - Sung Won Kim
PY  - 2026
DA  - 2026/10/30
JO  - PLOS One
DO  - 10.1371/journal.pone.0359774
UR  - https://doi.org/10.1371/journal.pone.0359774
AB  - UniHAR is a lightweight hierarchical multi-scale transformer based on wearable sensor data that has to be able to model time well; be able to recognize human activity across different datasets; and do so in real-time. There are several problems with what currently exists. UniHAR uses four major components: universal channel harmonization; adaptive multi-scale patch embedding; scale-adaptive Fourier cross-attention; factorized temporal-channel attention. The components allow for the ability to use various sensors and enables the low-complexity multi-scale fusion of features. UniHAR achieved an accuracy of 96.35%; 94.45%; 96.44%; and 99.23% on UCI-HAR, Opportunity, Daphnet, and PAMAP2 datasets, while being very lightweight and having high-speed inference capabilities; therefore, UniHAR is suitable for real-time healthcare applications.
SN  - 1932-6203
ER  -