Pavement Marking Incorporated With Binary Code for Accurate Localization of Autonomous Vehicles

Accurate localization is a critically important issue for autonomous vehicles as it closely related to the safety and efficiency of driving. However, current technologies vehicle face many challenges. To provide accurate robust services vehicles, we propose novel solution by employing newly designed pavement marking. This marking operates on color contrast, temperature binary code with some special features. We also trained customized an object detector based deep learning model: YOLOv5, integrated decoding algorithm. The system capable running at steady frame rate more than 50 FPS. Road trials up 80 km/h were conducted, satisfactory results confirmed feasibility robustness system. Specifically, common onboard camera, four continuous frames can be detected decoded correctly when speed slower 30 km/h. At least one higher (i.e., 30– km/h). With high-speed 18 even findings suggest that specially road associated algorithms viable economical option vehicles. performance has potentials further improvement using better hardware such faster CPUs, GPUs, thermal imaging techniques.