Detail publikačního výsledku

Deep Learning-Based Human Activity Classification with OTFS Radar and Attention Enhanced LSTM Networks

ALI, M.; MARŠÁLEK, R.

Originální název

Deep Learning-Based Human Activity Classification with OTFS Radar and Attention Enhanced LSTM Networks

Anglický název

Deep Learning-Based Human Activity Classification with OTFS Radar and Attention Enhanced LSTM Networks

Druh

Stať ve sborníku v databázi WoS či Scopus

Originální abstrakt

Radar-based human activity recognition (HAR) is emerging as a resilient alternative to camera systems in industrial environments, where occlusions, reflective surfaces, and privacy concerns limit vision-based methods. This paper presents a deep learning framework that integrates convolutional feature extraction, denoising, multi-head attention, and bidirectional LSTM layers to classify activities from OTFS radar delay–Doppler signatures. Experiments in cluttered industrial environments demonstrate that denoising improves robustness against clutter, while attention stabilizes performance across longer temporal sequences. With a sequence length of 50 frames and 32 attention heads, the proposed model achieves 95.2\% accuracy on unseen test data. Visualization using t-SNE confirms clear activity separation, with minor overlap between walking and multi-person walking due to shared Doppler patterns. These results highlight the effectiveness of combining OTFS radar with attention-based temporal modeling for reliable and efficient HAR in real-world industrial monitoring.

Anglický abstrakt

Radar-based human activity recognition (HAR) is emerging as a resilient alternative to camera systems in industrial environments, where occlusions, reflective surfaces, and privacy concerns limit vision-based methods. This paper presents a deep learning framework that integrates convolutional feature extraction, denoising, multi-head attention, and bidirectional LSTM layers to classify activities from OTFS radar delay–Doppler signatures. Experiments in cluttered industrial environments demonstrate that denoising improves robustness against clutter, while attention stabilizes performance across longer temporal sequences. With a sequence length of 50 frames and 32 attention heads, the proposed model achieves 95.2\% accuracy on unseen test data. Visualization using t-SNE confirms clear activity separation, with minor overlap between walking and multi-person walking due to shared Doppler patterns. These results highlight the effectiveness of combining OTFS radar with attention-based temporal modeling for reliable and efficient HAR in real-world industrial monitoring.

Klíčová slova

OTFS radar, human activity recognition, deep learning, attention mechanism, LSTM networks, integrated sensing and communication (ISAC), industrial monitoring, 6G sensing

Klíčová slova v angličtině

OTFS radar, human activity recognition, deep learning, attention mechanism, LSTM networks, integrated sensing and communication (ISAC), industrial monitoring, 6G sensing

Autoři

ALI, M.; MARŠÁLEK, R.

Rok RIV

2026

Vydáno

11.11.2025

Nakladatel

IEEE

Místo

Sofia, Bulgaria

ISBN

979-8-3315-9128-1

Kniha

2025 28th International Symposium on Wireless Personal Multimedia Communications (WPMC)

Strany počet

6

URL

https://ieeexplore.ieee.org/document/11351216

BibTex

@inproceedings{BUT199728,
  author="Malek Abdulmalek Ahmed {Ali} and Roman {Maršálek}",
  title="Deep Learning-Based Human Activity Classification with OTFS Radar and Attention Enhanced LSTM Networks",
  booktitle="2025 28th International Symposium on Wireless Personal Multimedia Communications (WPMC)",
  year="2025",
  pages="6",
  publisher="IEEE",
  address="Sofia, Bulgaria",
  doi="10.1109/WPMC67460.2025.11351216",
  isbn="979-8-3315-9128-1",
  url="https://ieeexplore.ieee.org/document/11351216"
}