MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541123380 A) filed by Merla Chandra Rao, Hyderabad, Telangana, on Dec. 8, 2025, for 'edge-computing architectures for ai inference in iot and smart devices.'
Inventor(s) include Mr. D. Shine Rajesh; Mr. Kammara Venkatarangaiah Achari; Mr Bala Krishna Varikuntla; Mr Enugurthi Gangaram; Dr. K. Manikandan; Mr. Santhosha B M; and Mr. Merla Chandra Rao.
The application for the patent was published on Jan. 2, under issue no. 01/2026.
According to the abstract released by the Intellectual Property India: "The rapid proliferation of intelligent IoT and smart devices demands power-efficient and scalable edge-computing architectures capable of performing real-time AI inference without reliance on cloud connectivity. This work presents an optimized edge-driven AI processing framework that integrates lightweight neural acceleration, adaptive caching, and event-driven computation to significantly enhance inference throughput under stringent energy constraints. The architecture employs hierarchical processing layers that dynamically distribute workloads between local nodes and gateway-level microservers using a context-aware offloading scheme. A model compression pipeline-incorporating quantization, structured pruning, and on-device incremental learning-reduces computational overhead while preserving accuracy for evolving datasets in dynamic environments. Hardware-software co-design enables integration with heterogeneous embedded platforms, supporting sensor fusion, predictive analytics, and secure federated learning. Experimental evaluation demonstrates substantial reductions in latency, bandwidth consumption, and power utilization compared to conventional cloud-assisted inference models. The proposed architecture provides a robust edge-intelligence foundation for future autonomous IoT ecosystems, including wearable healthcare systems, smart cities, environmental monitoring, industrial automation, and next-generation consumer electronics."
Disclaimer: Curated by HT Syndication.
