NSERC USRA Research Paper

Abstract

This research paper discusses the comprehensive exploration of adaptive and non-adaptive control systems utilizing Quanser's QCar platform. The study covers real-time computational capabilities, real-time computer vision techniques, and the educational potential of Quanser's QCar. The resulting systems – Adaptive Lane Following, Sign Detection, Traffic Light Detection, and Adaptive Cruise Control – showcase their effectiveness across diverse driving scenarios and environmental conditions. The research methodology involves distinct phases, starting with an Initial Onboarding Process. Familiarity with software platforms such as Simulink, ROS, Python, and more was essential. Understanding the QCar platform was facilitated through User Manuals and predeveloped Simulink documentation. In-depth studies in Autonomous Technology and Computer Vision informed system development. A structured approach encompassing Research, Planning, Development & Testing stages was followed. Research gathered insights from repositories, videos, research papers, and websites. This also involved creating final testing environments to simulate various lighting conditions which were called daytime, afternoon, and nighttime environments. This allowed for the QCar’s performance to be effectively evaluated and compared against each other. The Adaptive Lane Following system tackles fluctuating lighting conditions using adaptive thresholding for consistent lane tracking. The Sign Detection System identifies stop signs, serving as a foundation for more complex systems. The accurate Traffic Light Detection system enables safe navigation. The ACC and Object Detection system enhances QCar's safety in dynamic traffic. The study effectively combines theory with practical implementation, exploring both adaptive and non-adaptive control systems. The developed systems showcase capabilities across different scenarios. Additional potential research directions include Python threading, audio features, GPS Track Network, and ROS optimization. This research mainly contributes to educational autonomous vehicle innovation and development.