MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641051305 A) filed by A. Beno; Sofia; C. Peter Devadoss; J. Jenifer Grena; and M. Sharmila, Thoothukudi, Tamil Nadu, on April 22, for 'an intelligent edge-computing based ai-iot system for adaptive smart environment management.'
Inventor(s) include A. Beno; Sofia; C. Peter Devadoss; J. Jenifer Grena; and M. Sharmila.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "The present invention discloses an Intelligent Edge-Computing Based AI-IoT System for Adaptive Smart Environment Management. The system comprises a heterogeneous IoT sensor network layer; a plurality of intelligent edge computing nodes each equipped with a dedicated neural processing unit (NPU); an embedded AI inference engine hosting cooptimized neural network models including Convolutional LSTM occupancy prediction, Deep Reinforcement Learning environmental optimization, autoencoder-based anomaly detection, and federated user behavior learning models; an adaptive control module implementing hierarchical model predictive control; a cross-domain integration framework unifying HVAC, lighting, energy, security, and occupant comfort management; and a privacy-preserving federated learning module. The system performs real-time, low-latency AI inference locally at the network edge, eliminating dependence on centralized cloud servers for primary control operations. The aggregate inference latency of the full AI model ensemble is maintained below 50 milliseconds. The adaptive control module employs multi-objective Pareto-optimal optimization to simultaneously maximize occupant thermal comfort, indoor air quality, energy efficiency, and security. Federated learning with local differential privacy enables collaborative AI model improvement across distributed edge nodes without sharing raw sensor data, ensuring privacy preservation. The invention achieves significant improvements in energy efficiency, occupant comfort, response latency, and data privacy compared to conventional cloud-centric and rule-based smart environment management systems. The system is suitable for deployment across residential, commercial, industrial, healthcare, and urban infrastructure environments."
Disclaimer: Curated by HT Syndication.
