Copyright © 2026 Authors retain the copyright of this article. This article is an open access article distributed under the Creative Commons Attribution License which permits unrestricted use, distribution, and reproduction in any medium, provided the original work is properly cited.

@article{205126,
        author = {Aparna Tiwari and Siddhant Singh Kaushik and Shreya Chaudhary and Abhishek Kumar Yadav},
        title = {Hybrid Edge–TinyML Architecture for Intelligent and Energy-Efficient Industrial Automation Systems},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {13},
        number = {1},
        pages = {6220-6230},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=205126},
        abstract = {This research presents a Hybrid Edge–TinyML Architecture aimed at mitigating the significant issues of high latency, bandwidth saturation, and excessive energy consumption prevalent in conventional cloud-centric Industrial IoT (IIoT) frameworks. The system allows for reflexive, on-device anomaly detection by decentralizing intelligence and putting quantized 1D-Convolutional Neural Networks (1D-CNNs) directly on ESP32 microcontrollers.
The architecture uses a three-tier hierarchy—Extreme Edge, Local Gateway, and Global Cloud—to find a balance between short-term responsiveness and long- term analytical depth. Experimental results show that this hybrid method cuts latency by 93.7% and power use by 78% compared to streaming models. This framework is a strong, scalable, and energy-independent solution for the next generation of Industry 4.0 predictive maintenance systems. It maintains a classification accuracy of 94.6% even after 8-bit quantization.},
        keywords = {TinyML, Edge Computing, Industrial Internet of Things (IIoT), Anomaly Detection, Energy-Efficient Computing, Industry 4.0, Model Quantization, ESP32, Convolutional Neural Networks (CNN), Predictive Maintenance, Real-time Systems, Distributed Intelligence, Hardware-Software Co-design, On-device Inference, Low-latency Automation, Energy Harvesting, Signal Processing, Embedded AI, Digital Signal Processing (DSP), MQTT Protocol, Quantization-Aware Training (QAT).},
        month = {June},
        }