The Driverless team at QUTMS has been in development of an autonomous package for an electric racecar since 2020. Many simulations have been designed or adapted to develop perception and control algorithms while physical hardware was unavailable.
QEV-3, Lando, has been the first testbench for the driverless team and was used for hardware testing throughout 2021. As of November 2021, the team had successfully tested perception (LiDAR and camera) and control (steering actuation) on Lando. Additionally, the team took the autonomous package on QEV-3 to track the following month, testing these components with a safety driver to control throttle and brake only, while also gathering an abundance of real-world™️ data. This has allowed further development for real-time pipelines, using machine learning and clustering models to detect cones outlining the track circuit.
In 2022, the team was determined to race QEV-3 as one of the first autonomous racecars in Australia during the 2022 FSAE Australasian Competition. After months of hardware and software debugging, ensuring algorithms were reliable, and redesigning some critical systems, QEV-3 was able to drive itself in October. This was remote controlled at first, then operated under its own processes. Soon after, the team was on the road to the competition in December, where QEV-3 passed all technical inspections to be the second competition-ready FSAE Autonomous Vehicle.
As we begin 2023, the Driverless team is aiming to improve the reliability of navigation processes and streamline our software stack for faster and more effective development into the future.
The Driverless System is developed using the Robot Operating System (ROS2), for efficient data transfer and package management. We are constantly pushing the boundaries of cutting-edge to create a sophisticated and future-proof package for years to come.