Hardware and Software Development of a Small Scale Driverless Vehicle

Abstract

Formula Student Driverless is a competition that encourages engineering students to design autonomous racing vehicles to compete in international competitions. This paper proposes a software stack for an FSAE Driverless vehicle prototype. The software stack is divided into five subsystems. The perception subsystem detects the landmarks, the localization subsystem tracks the vehicle's position, the mapping subsystem creates a global map of landmarks observed, and the planning and controls subsystem decides how the vehicle navigates through the map. A LiDAR was used to detect waypoints on the map, and EKF-SLAM was implemented for mapping. PID and Pure Pursuit were utilized for control of the vehicle. This paper also discusses a scaled model for testing the said software stack. The design parameters of the small-scale prototype were replicated and scaled down from our own FSAE Electric Vehicle, Lemnos.