A Deep Learning Framework on Embedded ADAS Platform for Pedestrian Detection

This paper discusses the pedestrian detection and tracking system for real-time computer vision applications in autonomous vehicles and surveillance on the Nvidia Jetson AGX Xavier platform. The effectiveness of the system’s detection and processing speed are two evaluation metrics. The outcomes demonstrate its potential uses in autonomous cars, smart cities, and surveillance and set the stage for high-performance, real-time pedestrian tracking and detection in settings with limited resources. The pedestrian detection and tracking system, when combined with Nvidia Jetson Xavier AGX and a deep learning framework for pedestrian detection, can be thoroughly tested in a variety of urban road scenarios thanks to HIL testing. With the help of dSPACE Scalexio’s HIL validation, this method offers a dependable and realistic platform for thorough testing in a virtual setting.