MUMBAI, India, Jan. 9 -- Intellectual Property India has published a patent application (202521111153 A) filed by Symbiosis International, Pune, Maharashtra, on Nov. 13, 2025, for 'system and method for multi-stage traffic anomaly detection and congestion classification using multi-head attention architecture.'
Inventor(s) include Pratik Vinayak Jadhav; Dr. Deepali Vora; Dr. Abderrahim Benslimane; and Dr. Shruti Patil.
The application for the patent was published on Dec. 12, under issue no. 50/2025.
According to the abstract released by the Intellectual Property India: "A computer-implemented system for detecting traffic anomalies, classifying congestion levels, and identifying incident-caused anomalies in urban traffic networks using multi-head attention-based deep learning architecture is disclosed. The system comprises a processor coupled to a memory executing a data acquisition module collecting traffic sensor data integrated with weather and incident information, a data processing module performing temporal feature extraction and sliding window structuring, an anomaly identification module applying isolation forest and clustering algorithms, a multi-head attention LSTM module analyzing temporal dependencies through transformer blocks combined with recurrent layers, a multi-stage classification module predicting parallel outputs for anomaly detection, congestion categorization, and incident identification, and a natural language generation module producing human-readable textual descriptions through fine-tuned language models. The system provides technical advancement through reduced computational latency via parallel attention mechanisms, improved classification accuracy through hybrid neural network architectures, and enhanced interpretability through automated natural language report generation enabling real-time traffic management."
Disclaimer: Curated by HT Syndication.
