The goal to build RURacer-2 was to create improve platform performance, such as more accurate wheel encoders and more advanced embedded processors. Compared to RURacer-1, the new platform can reach four times lower latency (less than
Hardware and software architecture
The scaled vehicle chosen for this project is a similar 1/7th scale Traxxas XO-1. The XO-1 shown in Figure~\ref{fig:2-scaleVehicle} is highly capable of drifting as it features an optional rear-wheel drive system and can reach up to
We chose NVIDIA Jetson-TX2 to control the vehicle because of it's high computational power for real-time processing. TX2 is running the Linux kernel with an Ubuntu operating system created by NVIDIA called L4T. Also, TX2 features onboard WiFi, which we set up to connect to the router and the host computer for real-time communications. The host computer provides high computational power at the edge for resource-hungry algorithms such as motion planning or image processing. The communication framework is built on top of ROS to receive and send data from different processing nodes.
A microcontroller controls the steering wheel's servo and the speed of the brushless motor. Also, optical encoders are connected to the microcontroller, communicating to TX2 by Universal Asynchronous Receiver/Transmitter (UART) using \textit{rosserial}. Using UART, we achieve a
The IMU used for RURacer-2 is provided by Adafruit-BNO055, providing 9 DOF sensing information, that it has been difficult before. We used a 3.2 Mega Pixels (MP) FLIR Blackfly GigE camera at