MUMBAI, India, May 1 -- Intellectual Property India has published a patent application (202641050173 A) filed by Sudhakar Hallur; Amruta Kulkarni; Anil Gavade; and Priyanka Gavade, Shindolli, Karnataka, on April 20, for 'a method for multimodal driver emotion detection and risk prediction using adaptive attention fusion.'
Inventor(s) include Sudhakar Hallur; Amruta Kulkarni; Anil Gavade; and Priyanka Gavade.
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 discloses a Method For Multi-Modal Driver Emotion Detection And Risk Prediction Using Adaptive Attention Fusion. The method has Data Acquisition Layer which comprises of Vision Sensing Module (201), Physiological Sensing Module (206), Audio Sensing Module (211), Vehicle Telemetry Interface (213), and Environmental Sensing Module (216). The method is configured to acquire driver-related data through said sensing modules, performs Sensor Reliability Estimation, Data Preprocessing, and modality specific Feature Extraction through a multi-architecture neural network. Extracted features are dynamically fused through an Adaptive Attention Fusion Layer (105) so as to generate a unified fused feature. Based on this fused feature, an Emotion Classification Engine Layer (107) determines driver emotional states and an Emotion-to-Risk Mapping Layer (108) computes quantifiable driving risk scores enabling proactive ADAS-integrated safety interventions, with all biometric data processed locally within the embodiment."
Disclaimer: Curated by HT Syndication.
