MUMBAI, India, Sept. 22 -- Intellectual Property India has published a patent application (202641109050 A) filed by Ch Venkata S S P Kumar; Dr Sachin Choudhari; Krutika Kushal Lohiya; N Bhuvaneswari; Dr. G. Vijaya; Prashant Govardhan; Bhagyashree C. Hambarde; Dr. Saurabh Pandey; and Dr. B. Aswini on September 11, 2026, for System And Method For Real-Time Multimodal Sensor Fusion Using Graph Neural Networks.
Inventors include Ch Venkata S S P Kumar; Dr Sachin Choudhari; Krutika Kushal Lohiya; N Bhuvaneswari; Dr. G. Vijaya; Prashant Govardhan; Bhagyashree C. Hambarde; Dr. Saurabh Pandey; and Dr. B. Aswini.
The application for the patent was published on September 18, 2026, under issue no. 38/2026.
Abstract: A system and method for real-time multimodal sensor fusion using graph neural networks comprises heterogeneous sensor units configured to acquire image data, ranging data, and inertial data, signal conditioning circuits configured to filter and digitize sensor outputs, and a time synchronization unit configured to associate measurements with a common time reference. A preprocessing processor generates modality-specific features, and a graph construction processor generates a dynamic multimodal graph having nodes representing the features and edges representing spatial, temporal, and cross-modal relationships. A graph neural network processor performs iterative message passing between connected nodes and generates updated node representations. A fusion processor combines the updated node representations to generate a fused representation of an object, physical condition, or environmental state, and an output processor generates a real-time fused output for an external hardware device.
Disclaimer: Curated by HT Syndication.
