EUVIMEDEUVIMED entdecken
Open Access · CC BY 4.0

AI-Driven Signal Integrity and Electromagnetic Interference Prediction Using Physics-Informed Neural Networks and Explainable Edge Intelligence

Balachandra Kumaraswamy

Journal of Intelligent Decision Making and Information Science · 03.10.2026 · 3 (15s) · 2304-2319

DOI: 10.17762/jidmis.v3.4812 ↗

Kurz gefasst

Fachartikel in der Zeitschrift Journal of Intelligent Decision Making and Information Science, online erschienen am 03.10.2026. Verfasst von Balachandra Kumaraswamy. Der Volltext ist frei zugänglich (Open Access, CC BY 4.0).

AbstractBibliografische AngabenZitierenVerwandte Studien

Abstract (Originalsprache)

High-speed electronic systems become more complex and operate more frequently, it is becoming more critical to ensure their reliable operation with accurate signal integrity (SI) prediction and electromagnetic interference (EMI) analysis and prediction. In this study, an AI-based framework that combines Physics-Informed Neural Networks (PINNs), Explainable Artificial Intelligence (XAI) and lightweight edge intelligence is proposed to effectively and accurately predict the EMI for the aim of achieving efficient design. A set of communication parameters and electromagnetic characteristics like the RSSI, SNR, frequency, transmission power, distance and the characteristics of the packet are preprocessed and then ANN, CNN, DNN and PINN models are trained with those parameters. Physics-informed constraints ensure consistency in predictions and parameters are identified that are influential by using SHAP. The results of the experiments show that PINN has the lowest MAE (1.9), the lowest RMSE (2.6) and an R² of 0.97. The feature that explains the most is frequency, with a SHAP importance of 0.31. Moreover, PINN-Lite can reach the inference latency of 6.1ms and model size of 1.3MB. The framework is thus accurate, interpretable and computationally efficient to enable EMI prediction for intelligent PCB optimization and edge-based monitoring.

Zitieren

APA

Kumaraswamy, B. (2026). AI-Driven Signal Integrity and Electromagnetic Interference Prediction Using Physics-Informed Neural Networks and Explainable Edge Intelligence. Journal of Intelligent Decision Making and Information Science, 3(15s), 2304-2319. https://doi.org/10.17762/jidmis.v3.4812

Vancouver

Kumaraswamy B. AI-Driven Signal Integrity and Electromagnetic Interference Prediction Using Physics-Informed Neural Networks and Explainable Edge Intelligence. Journal of Intelligent Decision Making and Information Science. 2026;3(15s):2304-2319. doi:10.17762/jidmis.v3.4812

BibTeX
@article{Kumaraswamy2026,
  author = {Kumaraswamy, Balachandra},
  title = {AI-Driven Signal Integrity and Electromagnetic Interference Prediction Using Physics-Informed Neural Networks and Explainable Edge Intelligence},
  journal = {Journal of Intelligent Decision Making and Information Science},
  year = {2026},
  volume = {3},
  number = {15s},
  pages = {2304--2319},
  doi = {10.17762/jidmis.v3.4812}
}
Kostenlos starten →