Self-driving car

Full disclosure: This car is perhaps not quite as big as the car you envisioned when you read the headline.

img_8988 (1)


Zheng Wang from Bridgwater State University has used a Raspberry Pi and some other hardware to modify a remote-controlled (RC) car to follow a track, detect, understand and respond to stop signs and traffic lights, and to avoid collisions. Once scaled up and able to do parallel parking, you’ve got something that looks a bit like Google’s self-driving car project. (A bit.)

Here’s a rather neat technology demo.

OpenCV Python Neural Network Autonomous RC Car

Bridgewater State University COMP 502 Project, May 2015 Self driving RC car: OpenCV neural network – Steering Haar-cascade classifiers – Stop sign and traffic light detection Ultrasonic sensor – Front collision avoidance Raspberry Pi – Data streaming (video and sensor) Arduino – RC car control

So what’s happening here? The Pi is hooked up to a Raspberry Pi Camera Module and an ultrasonic sensor. Two client programs on the Pi are used to serve the information it gathers from those devices to another computer over WiFi, with streaming video. The RC controller for the car is given instructions by an Arduino which is hooked up to the computer doing the processing by USB.

Zheng has provided a very detailed writeup, which dives into the maths behind all of this, and provides a look at the neural network on the machine doing the processing.


Geometric model for detecting distance with monocular vision

Head over to his website to have a look – it’s a fascinating read. Thanks Zheng – drop us a line if you take this project any further!