MUMBAI, India, Aug. 10 -- Intellectual Property India has published a patent application (202641089490 A) filed by Dr. Anand M.; Dr. Aishwarya N. Kumar; Mrs. Rummana Firdaus; Mrs. Geetha A. L.; Mrs. Meghana A.; Mrs. Shilpashri V. N.; Mrs. Poojitha G. S.; Dr. Thilagavathy R.; Mrs. Seema Firdose; Dr. Sheshadri S. N.; Dr. Pradeep Kumar R.; and Mrs. Syeda Nausheen Fathima on July 22, 2026, for A Hardware-Assisted System And Method For Agricultural Commodity Price Prediction Using Supervised Data Fusion.
Inventors include Dr. Anand M.; Dr. Aishwarya N. Kumar; Mrs. Rummana Firdaus; Mrs. Geetha A. L.; Mrs. Meghana A.; Mrs. Shilpashri V. N.; Mrs. Poojitha G. S.; Dr. Thilagavathy R.; Mrs. Seema Firdose; Dr. Sheshadri S. N.; Dr. Pradeep Kumar R.; and Mrs. Syeda Nausheen Fathima.
The application for the patent was published on July 31, 2026, under issue no. 31/2026.
Abstract: The invention provides a distributed system (100) that acquires physical production data from field sensor nodes (112), storage or transit data from supply-chain sensor nodes (114), and verified market and weather data through interfaces (116, 118). An edge gateway (120) verifies integrity, aligns timestamps and generates compact commodity-state packets. A reliability stage (130) computes source-specific quality and forms weighted features. A supervised model engine (140) combines a shared base regressor with a commodity-region-horizon residual model. An uncertainty estimator and confidence gate (150) release a predicted price, interval and direction only when stored acceptance criteria are met. Accepted outputs can drive a bounded infrastructure-control interface (170). Verified realised prices are reconciled as labels (180), and detected residual or feature drift triggers validated retraining (190). FIG. 1 may accompany the abstract.
Disclaimer: Curated by HT Syndication.
