Autonomous vehicles (AVs) are becoming increasingly important, but current research is hindered by expensive platforms that lack real-world simulatability and affordability. To address this challenge, we present the "Go-Kart AV Platform," an open-source, modular, electric vehicle designed for academic research. By leveraging a smaller platform with reduced costs and increased scalability, our design enables more diverse teams to investigate safety and performance limitations in realistic scenarios.
The Go-Kart AV Platform is a one-third scale autonomous electric go-kart with open-source mechatronics design and fully functional autonomous capabilities. Its smaller size makes it more affordable for academic departments and research groups, while still providing sufficient similitude in dynamics, control, and drivability compared to full-scale vehicles. This platform offers immense value to universities and research institutions by fostering collaboration towards the open development and validation of AVs.
To further improve the platform’s mechatronic, sensing, and software systems, future work will focus on continuous improvement and exploring new features like imitation learning algorithms that enable dynamic cooperative control between drivers and vehicles. By leveraging different driving modes and human-machine interactions, we can advance research in autonomous transportation and promote innovation in the field.
In summary, the Go-Kart AV Platform offers a valuable solution for academic researchers to investigate safety and performance limitations of autonomous vehicles without breaking the bank. Its open-source design enables collaboration, continuous improvement, and innovation in the rapidly evolving field of autonomous transportation.
Autonomous Racing: Developing High-Performance Software and Hardware for Unprecedented Levels of Independence
