EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

Knowledge Distillation and Quantization-Aware Compression for Remaining Useful Life Prediction on Resource-Constrained IoT Devices

Kurnianingsih, Nurseno Bayu Aji, Dwiana Hendrawati, Eko Susanto, Retno Widyowati, Helmy Yusuf, Putut Har Riyadi, Neny Purwitasari, Heli Siti Halimatul Munawaroh, Sukardiman

Engineering, Technology & Applied Science Research · 11.08.2026 · 16 (4) · 37784-37792

DOI: 10.48084/etasr.19973 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Engineering, Technology & Applied Science Research, online erschienen am 11.08.2026. Verfasst von Kurnianingsih und 9 weiteren Autor:innen. Die Arbeit nennt 30 Literaturangaben. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

Models that achieve the best desktop accuracy in Remaining Useful Life (RUL) prediction do not necessarily meet the deployment constraints on IoT microcontrollers. The dominant evaluation paradigm rewards offline accuracy, yet condition-based maintenance ultimately requires inference within strict memory and latency budgets. This paper addresses this gap by evaluating three composable compression mechanisms, namely matched-capacity architecture scaling, Knowledge Distillation (KD), and INT8 Quantization-Aware Training (QAT), on the C-MAPSS FD001 benchmark with ESP32-S3 hardware validation. The analysis shows that the temperature-scaled KD is mathematically inert for scalar regression: KD improves cross-seed consistency (3.5× variance reduction) more than mean accuracy. Matched-capacity scaling compresses a 147K-parameter TTSNet teacher into a 33K StudentTTSNet (77.7% reduction), and INT8 QAT adds a 3.98× weight-level reduction, yielding a 70.9 KB deployed model at only 1.6% RMSE cost. On the ESP32-S3 microcontroller, the best offline GRU baseline fails to produce correct predictions in the tested TFLite Micro configuration due to the 1024-tensor allocation limit, whereas the compressed student model runs in real time at 56.8 ms (FP32) and 83.9 ms (INT8). Preliminary cross-domain experiments on a small plant-extraction dataset suggest that the same pipeline can regularize smaller data settings, indicating that runtime compatibility deserves validation as early as offline accuracy in IoT-bound RUL research.

Zitieren

APA

Kurnianingsih, Aji, N. B., Hendrawati, D., Susanto, E., Widyowati, R., Yusuf, H., Riyadi, P. H., Purwitasari, N., Munawaroh, H. S. H., & Sukardiman (2026). Knowledge Distillation and Quantization-Aware Compression for Remaining Useful Life Prediction on Resource-Constrained IoT Devices. Engineering, Technology & Applied Science Research, 16(4), 37784-37792. https://doi.org/10.48084/etasr.19973

Vancouver

Kurnianingsih, Aji NB, Hendrawati D, Susanto E, Widyowati R, Yusuf H, et al. Knowledge Distillation and Quantization-Aware Compression for Remaining Useful Life Prediction on Resource-Constrained IoT Devices. Engineering, Technology & Applied Science Research. 2026;16(4):37784-37792. doi:10.48084/etasr.19973

BibTeX
@article{Kurnianingsih2026,
  author = {Kurnianingsih and Aji, Nurseno Bayu and Hendrawati, Dwiana and Susanto, Eko and Widyowati, Retno and Yusuf, Helmy and Riyadi, Putut Har and Purwitasari, Neny and Munawaroh, Heli Siti Halimatul and Sukardiman},
  title = {Knowledge Distillation and Quantization-Aware Compression for Remaining Useful Life Prediction on Resource-Constrained IoT Devices},
  journal = {Engineering, Technology & Applied Science Research},
  year = {2026},
  volume = {16},
  number = {4},
  pages = {37784--37792},
  doi = {10.48084/etasr.19973}
}
Kostenlos starten →