MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641049904 A) filed by Easwari Engineering College; and Srm Institue Of Science And Technology, Ramapuram Campus, Chennai, Tamil Nadu, on April 20, for 'adaptive traffic flow forecasting using ga-bo optimized spatio-temporal graph neural networks.'
Inventor(s) include Bala Shivani P. D; Kaviya R. V; and Muhemin Sheriff M.
The application for the patent was published on May 1, under issue no. 18/2026.
According to the abstract released by the Intellectual Property India: "Urban traffic congestion remains a critical challenge in modern cities, leading to increased travel time, fuel consumption, and air pollution. Classical traffic forecasting methods fail to accurately capture complex spatio-temporal dependencies and adapt dynamically to time-varying traffic conditions, resulting in limited accuracy and scalability. This work proposes a novel GA- BO-STGNN framework integrating a Spatio-Temporal Graph Neural Network (ST-GNN) with a hybrid Genetic Algorithm-Bayesian Optimization (GA-BO) method. The system processes real-time traffic data from sensors and cameras, effectively capturing spatial interactions across road networks and temporal traffic variations. The GA-BO optimizer dynamically tunes network hyperparameters including learning rate, hidden layers, and batch size, improving prediction accuracy and convergence speed. Developed entirely in MATLAB, the framework demonstrates strong adaptability, robustness, and accuracy in forecasting traffic flow under uncertain conditions. The key novelty lies in simultaneously optimizing deep network architecture and hyperparameters, enabling scalable and intelligent smart city traffic control."
Disclaimer: Curated by HT Syndication.
