Data Acquisition - Simulating the Racetrack

March 16, 2022

The Driving Simulator project is an end-to-end system that includes everything from a cockpit to data visualisation and analysis in Redback Cloud. The team utilises Assetto Corsa for the simulation engine as it provides exceptionally good value for how accurate it is. 3D models of Redback’s cars and tracks are built by importing the CAD models whilst continuing to use track maps from previous competitions. Detailed physics models of each of the cars are also built with the help of experienced vehicle dynamicists on the team.

POV shot of RB21-E in Assetto Corsa
POV shot of RB21-E in Assetto Corsa

Assetto Corsa supports custom in-game app plugins written in python, thus allowing us to build a telemetry app that collects all of the vehicle data from the simulation and gather it into Redback Cloud. Assetto Corsa provides over 150 data points from the car, which is far greater than we can achieve from real cars. The Driver Simulation team uses the same concepts as the firmware for integrating Redback Cloud, such as the WebSocket connection to Redback Cloud for passing control messages from the UI and a UDP stream for live telemetry. This means we can make full use of our live telemetry and data analysis features with the driving simulator which improves the accuracy of the simulator models, the design of the car and the driver performance.

RB21-E on track in the simulator, allowing drivers to practice accuracy and performance in the car

Redback is also working on a cockpit driving simulator to match the driving position, wheel and pedal experience of the real car as best as possible. The team has designed a custom ergonomics rig and brake pedal system based on load cells which we are set to build throughout 2021. Combining the ergonomics rig and driving simulator has the potential to complement each other greatly by allowing us to test the ergonomics of potential driving positions while operating the wheel and pedals to measure specific driver performance in each potential design. The team can further adjust the simulator rig to match the driving position of the actual car as the design changes each year.

CAD Render of Redback Racing’s current simulator

Currently, the team owns a Logitech G27 wheel and pedal set which is functional but limited on the simulation of the feel of our cars and is the reason why the team has designed our own brake pedal. We are looking out for any potential sponsors to assist and work alongside the Driver Simulation/DAQ team to achieve a far more immersive simulator experience to further train our driver’s performers whilst at competitions on the track. Improved drivers not only means we perform better on the day at our competition, but it also means laps at testing sessions are more consistent and higher quality data is produced. The drivers can also provide better feedback on the performance and feel of the car.

2022 will be the first year Redback uses the driving simulator in the FSAE-A competition where we aim to load the track into the sim the day before so our drivers are much more familiar with the track when going into the Auto-X and Endurance events.

Lap Simulator

The lap simulator is another project born out of the lockdowns in 2020 by a small team made of up of vehicle dynamics and software team members. There are 2 major components of the lap simulator project:

  • the first being the steady-state point mass simulator engine
  • the second is integration with Redback Cloud including service in the backend and user interface.

The point mass simulator library can either be used through our command-line application or by an application that connects with Redback Cloud. Engineers design the car graphically, forward it over to the simulator for processing, and view the results of the simulations in our data analysis system.

Redback Cloud displays the DAQ team the results from the simulator

The system is designed so that multiple lap simulators are able to be connected at the same time whilst requests can be load balanced between them. Simulators can either be deployed in the Cloud where autoscaling is frequently used, locally on engineers’ machines for testing code changes or on high-performance computing clusters for optimal performance.