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{192772,
        author = {Dr M Chandrashekar and Nandini and Nikhil},
        title = {A Lightweight Machine Learning–Driven Adaptive Temperature Control System for Real-Time Embedded Power Modulation},
        journal = {International Journal of Innovative Research in Technology},
        year = {2026},
        volume = {12},
        number = {9},
        pages = {4748-4752},
        issn = {2349-6002},
        url = {https://ijirt.org/article?manuscript=192772},
        abstract = {This paper presents a deployable TinyML-driven adaptive temperature control system for real-time embedded power modulation on resource-constrained microcontrollers. Conventional PID controllers require manual tuning and exhibit performance degradation under nonlinear and time-varying thermal dynamics. The proposed framework replaces fixed control laws with a quantized shallow neural network that directly approximates nonlinear thermal mappings while preserving strict real-time constraints. A Lyapunov-based ultimate boundedness analysis is formally derived to guarantee closed-loop stability under bounded neural output and actuator saturation. Experimental and simulation results demonstrate a 43.7% reduction in RMSE, 35% faster settling time, 64% reduction in overshoot, and 11.2% energy savings compared to classical PID control. Embedded profiling confirms 1.8 ms inference latency and 28 KB RAM usage on a Cortex-M0+ platform, validating practical feasibility. The proposed architecture establishes a scalable and computationally efficient pathway toward intelligent edge-level adaptive control in embedded power systems..},
        keywords = {TinyML, Embedded Control, Adaptive Systems, Neural Network Control, PWM Modulation, Edge AI.},
        month = {February},
        }