From Simulation to the Race Track: Development, Testing, and Deployment of Autonomous Racing Software

The development and testing of autonomous racing technology is a complex and challenging process that requires significant resources, expertise, and innovation. In this paper, the team Autonomous Racing Graz provides insights into their development and testing systems, which were refined during the participation in two ROBORACE seasons. We introduce the autonomous racing stack used by the team, and provide an overview of the modules used for system analysis, testing, and simulation. Our focus is on presenting solutions for efficient data analysis and comprehensive testing in simulation, as well as ways to achieve fast and reliable remote deployment in a time-constrained scenario. All systems have been designed with the goal of optimizing ease-of-use, development speed, and reducing the number of errors. Overall the paper aims to enable small teams to develop, test and deploy robust autonomous vehicle software with minimal resources.