MUMBAI, India, Jan. 2 -- Intellectual Property India has published a patent application (202541122776 A) filed by B V Raju Institute Of Technology, Narsapur, Telangana, on Dec. 5, 2025, for 'a hybrid spatial-temporal deep learning system for automated traffic accident identification and rapid response triggering.'
Inventor(s) include R. Neelima; Mastan Mohammed Meera Durga; Nagaram Ramesh; Sriharsha Vikruthi; Indumathi V; J. Manikandan; and Kenchali Shivaji Rao.
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: "Road accidents are among the top causes of mortality worldwide, with delayed detection and reporting contributing significantly to loss of life. Traditional surveillance systems are insufficient for real-time accident recognition and notification. This paper proposes a real-time, automated accident detection and emergency alert system that leverages existing CCTV infrastructure and advanced deep learning models. The system integrates YOLOv5 (You Only Look Once version 5) for realtime object detection and an LSTM (Long Short-Term Memory) network for temporal motion analysis. Anomalous behaviors such as sudden stops or collisions are identified, prompting an automated alert system that communicates essential information-including timestamp, visual evidence, and GPS location-to emergency responders via SMS, email, or API. This solution demonstrates high detection accuracy, low false positives, and fast alert response in varied environments. The system is scalable and adaptable for deployment in smart cities to enhance public safety and emergency response capabilities."
Disclaimer: Curated by HT Syndication.
