MUMBAI, India, Feb. 6 -- Intellectual Property India has published a patent application (202511131981 A) filed by Panipat Institute Of Engineering And Technology, Samalkha, Haryana, on Dec. 26, 2025, for 'adaptive multi-horizon energy demand forecasting framework integrating real-time data fusion, hybrid machine-learning prediction, and autonomous reinforcement-based model optimization.'
Inventor(s) include Prof. Poonam Panwar.
The application for the patent was published on Feb. 6, under issue no. 06/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses an adaptive multi-horizon energy-demand forecasting system that integrates real-time heterogeneous data acquisition, automated preprocessing and feature engineering, hybrid machine-learning and statistical prediction models, and an autonomous feedback controller supported by reinforcement learning. The system processes weather data, IoT-meter readings, industrial load measurements, grid-sensor inputs, renewable-generation profiles, and historical datasets to generate high-quality predictive features used by a coordinated deep-learning and statistical forecasting engine. A dynamic weighted fusion mechanism combines predictive outputs, while the adaptive controller continuously evaluates accuracy, updates parameters, and performs selective retraining to counter concept drift and fluctuating consumption patterns. A multi-horizon prediction layer delivers deterministic forecasts, probabilistic intervals, and anomaly indicators across time ranges from seconds to months. The invention provides a unified, scalable, and self-optimizing forecasting framework suitable for utility grids, industrial energy systems, renewable integration, and smart-city operations."
Disclaimer: Curated by HT Syndication.
