MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541123225 A) filed by Vellore Institute Of Technology, Vellore, Tamil Nadu, on Dec. 6, 2025, for 'neural network system for tropical cyclone wind-pressure relationship estimation.'
Inventor(s) include Dr. Jayakumar K; Sohum Chatterjee; Mantram Dwivedi; and Kagwade Jai Sachin.
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 present disclosure provides a neural network system for estimating wind-pressure relationships in tropical cyclones. The system includes a data preprocessing module (1) configured to receive tropical cyclone meteorological data and to normalize the tropical cyclone meteorological data using min-max scaling to generate normalized input features, a sequence generation module configured to convert the normalized input features into sequential data having a predetermined sequence length for temporal analysis, a recurrent neural network module including at least one of a Long Short-Term Memory (LSTM) model (5) and a Gated Recurrent Unit (GRU) model (6), a training module configured to train the recurrent neural network module using historical tropical cyclone data, and a prediction module configured to generate wind speed estimates for tropical cyclones undergoing eyewall replacement cycles based on the trained recurrent neural network module."
Disclaimer: Curated by HT Syndication.
