Pedestrian Safety

Pedestrian safety research aims to reduce pedestrian injuries and fatalities through improved understanding and prediction of pedestrian-vehicle interactions and environmental factors. Current research heavily utilizes computer vision, employing convolutional neural networks (CNNs) like VGG-19 and ResNet50, and deep learning models such as TabNet, to analyze video and image data for pedestrian detection, activity recognition, and risk assessment. These advancements, coupled with the use of ultra-wideband (UWB) radar and automated machine learning (AutoML) techniques, are leading to the development of real-time safety systems and data-driven strategies for improving infrastructure and driver/pedestrian behavior. The ultimate goal is to create safer road environments through proactive risk mitigation and improved situational awareness.