Prime Directive Solution in Spring 2024 Simulation Racing Series
Written by Virgile Foussereau (EECS) and Quang Huynh (ME), UC Berkeley students. Please download the PDF file to view their paper.
Abstract—This paper presents the record-breaking solution developed by the team Laplace Racing for the Robot Open Autonomous Racing (ROAR) Simulation Racing Series. Our solution outperforms all previous competitors in the autonomous racing domain by leveraging a model-free combination of Pure Pursuit for lateral control and a PID controller for longitudinal dynamics. Emulating the racing vehicle in the CARLA Research Environment, our approach focuses on achieving maximum speed and efficiency on the Monza track. We detail the architecture of our winning controller, but also the different approaches we have taken and their potential. Furthermore, we provide insights into our methodologies and optimization techniques that contributed to our success in the competition. This paper serves as a comprehensive analysis of our winning solution, and discusses potential future developments for further enhancing autonomous racing performance.