Abstract

The Pi Car initiative aims to build a compact vehicle platform for future testing of autonomous driving algorithms. Inspired by the Formula Pi Kickstarter project, we worked as part of a larger team under Drs. Humberto Gonzalez and Sylvia Zhang to produce mechanical, electrical, and software-side designs for a Raspberry Pi powered car. As the software communications team, we specifically developed three key interactions with the Raspberry Pi: (1) positional data feeds from motion-capture cameras, (2) input reads from the vehicle’s on-board sensors, and (3) the scheduling of these data inputs into external applications (e.g. web GUIs, controller programs).

We designed and tested programs written in Python to collect velocity and accelerometer metrics from a prototype’s IMU and encoders. Similarly, we used the Twisted-Python networking framework to process and distribute position coordinates from Dr. Gonzalez’s infrared camera system. With the completion of this project, we have provided a foundation for data communications of a single Pi Car vehicle, which can be further extended as the Pi Car initiative scales.