MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641051703 A) filed by Dayananda Sagar University, Bengaluru, Karnataka, on April 23, for 'a system and method for personalized real-time seizure prediction and adaptive alert generation using deep neuro-fuzzy reinforcement learning on edge devices.'
Inventor(s) include Diana George; Dharmendra D. P; Manoj U; Madan B R; Rakshith S Reddy; Mithun Gowda; Bharath M B; Deepthi Chamkur V; and Dr. Santosh Kumar J.
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: "The present invention relates to a system and method for real-time seizure prediction and adaptive alert generation using electroencephalogram (EEG) signals, integrating deep learning, fuzzy logic, and reinforcement learning techniques. The system comprises a deep neural network-based feature extraction module configured to process multi-channel EEG signals and generate temporal representations of preictal patterns. A fuzzy inference module is employed to transform prediction outputs into confidence-aware risk levels using linguistic variables, thereby addressing uncertainty in seizure prediction. A reinforcement learning-based decision module, implemented using a Deep Q-Network (DQN), dynamically optimizes alert generation policies based on patient-specific feedback, seizure occurrence history, and alert fatigue. The system is further configured for deployment on edge computing devices to enable low-latency, privacy-preserving operation. The proposed invention reduces false alarm rates while maintaining high sensitivity through adaptive decision-making, thereby enhancing reliability and personalization in seizure warning systems. The invention is applicable in wearable healthcare devices and real-time neurological monitoring systems."
Disclaimer: Curated by HT Syndication.
