A computer vision controlled autonomous vehicle for QUT’s Droid Racing Challenge (DRC) challenge.
In our first year (2018) the car had the 3rd fastest time against the National competitors and Fastest for Queensland competitors.
For more information about the DRC, visit QUT Robotics Club webpage: https://qutrobotics.com/droid-racing-challenge/
Specifications & Techniques used
Raspberry Pi Nano.
Raspberry Pi Camera V2.1
Sparkfun 16 channel servo hat
Raspberry Pi 3B+, Linux Raspbian, OpenCV.
Losi r/c car
Soldering, Assembly, perspex folding.
Servo & ESC Calibration.
Success follows frustration.
Have Programming and Mechanical Failsafe’s.
Make vision calibration easy.
Good front end user control is necessary.
Test multiple theories.
Build it big and make it work, then refine later.
It takes 5 revisions to reach the final solution. Be patient and test and learn from failures.
How to discover what you don’t know.
Correctly identifying failure.
Learning to work with redundant hardware.
Vision controlled via Raspberry Pi3B+.
Pi Racer has reached it’s computatioanl and physical limits. Pi Racer has been upgraded to “SmartyKart” and upgraded again to “The Red Fox”.