MUMBAI, India, Nov. 2 -- Intellectual Property India has published a patent application (202641124683 A) filed by B Jyothi; and Koneru Lakshmaiah Education Foundation on October 17, 2026, for Environment-Guided Reliability Modelling For Multisensory Pedestrian Perception In Autonomous Vehicles..
Inventors include B Jyothi; P. Thanmai Saran; and P. Bhavana.
The application for the patent was published on October 30, 2026, under issue no. 44/2026.
Abstract: Nowadays,adaptive multimodal fusion system for robust pedestrian perception in autonomous vehicles under challenging environmental and weather conditions. Pedestrian detection becomes difficult during low-light, night time, rainy, hazy, and poor-visibility conditions because individual sensors may not provide reliable information in every situation. To address this limitation, the proposed system combines information from an RGB camera, thermal sensor, and 4D radar orange -based sensor to improve pedestrian detection and enhance the robustness of the vehicle perception system. The proposed approach integrates Artificial Intelligence, Computer Vision, Deep Learning, and Multimodal Sensor Fusion to collect and combine complementary information from multiple sensors. The key idea of the architecture is to estimate the confidence and effectiveness of each sensor according to the current environmental conditions and dynamically assign different weights rather than giving equal importance to all sensors. The fused information is then provided to a deep-learning-based detection model to identify pedestrians more reliably in challenging conditions. Compared with individual sensor-based detection and conventional fusion approaches, the proposed adaptive fusion method aims to reduce false detections and improve pedestrian detection performance for autonomous vehicles. The proposed framework can be further extended to support collision-risk prediction, automatic emergency braking, intelligent steering,
Disclaimer: Curated by HT Syndication.
