MUMBAI, India, June 16 -- Intellectual Property India has published a patent application (202611053622 A) filed by Indian Institute Of Technology, Rupnagar, Punjab, on April 27, for 'a resource-aware embedded artificial intelligence system for low-resource iot devices.'
Inventor(s) include Dr Santosh K Vipaprthi; Dr Ashwani Sharma; Komal; and Dishang Dalwadi.
The application for the patent was published on June 5, under issue no. 23/2026.
According to the abstract released by the Intellectual Property India: "The present invention relates to a resource-aware embedded artificial intelligence system (100) for performing neural network inference on low-resource Internet of Things devices. The system comprises an input acquisition module, a preprocessing module configured to generate a compact tensor representation, a communication module, an embedded inference runtime executing a quantized neural network model, a decision logic module, and an output module. In a hybrid configuration (200), preprocessing is performed on a host device and a compact tensor is transmitted to the embedded device. The system utilizes reduced-precision model representation and a bounded-memory runtime for execution within constrained hardware environments. A structured tensor-transfer mechanism and a power optimization module enable efficient operation. The system supports multiple input modalities and provides structured output generation through embedded processing."
Disclaimer: Curated by HT Syndication.
