A Location-Based Pedestrian Behavior Recognition Platform for Enhanced Traffic Safety Using Dual Logic and V2P Technology

Abstract

This paper proposes a location-based pedestrian behavior recognition platform aimed at enhancing traffic safety, particularly for vulnerable road users. Utilizing accelerometer and GPS sensors embedded in smartphones and wearable devices, the platform employs a dual logic system to detect falls and assess pre- and post-fall scenarios in real-time. The proposed system filters, processes, and analyzes walking behavior to identify critical events such as sudden movements and falls, thereby supporting emergency alerts and traffic signal interventions. Integrating Vehicle-to-Pedestrian (V2P) communication technology and a fuzzy logic-based fall cognition model, the platform offers predictive walking route guidance. Additionally, a novel grid-based route modeling system is introduced for efficient pedestrian path prediction, minimizing infrastructure requirements. Data handling is structured across multiple layers including BigData processing, context awareness, and knowledge delivery, enabling personalized and scalable services. By bridging human behavioral patterns with traffic safety infrastructure, this platform supports safer pedestrian mobility, real-time behavior inference, and proactive traffic management, ultimately contributing to the development of smart and human-centric urban environments.