Eyes on the road

Patroklos Anagnostou, Russell Buchanan, Johan Byther, Johnathan Lehner

Eyes on the road is a hardware and software sensor suite designed to prevent drivers from falling asleep at the wheel. Using advanced sensor fusion and face tracking trained with a neural network we not only can tell when a driver is becoming fatigued but also predict tiring situations. We use a Bosch XDK and Logitec webcam to monitor car swerving, temperature, driver nodding off and more.

Sleeing at the wheel kills ...

Every year, drivers nod off, then veer off into oncoming traffic.

You have to act on the symptoms

Tiredness is treacherous. You get too tired to stop before you get too tired to continue. If you are incapable of deciding to take a break, you are a danger to yourself and others. But people forget.

Early warning is vital

We look at the position of the face and eyes, sudden movements indicative of the driver entering micro-sleep, environmental factors such as humidity and temperature, and whether you are driving at night or day. By combining the information from diverse systems, we can increase the accuracy beyond what one single component can provide.

To run

Checkout and unzip the data files. Ask us for a pickle file of net6 (or compute your own by running "kfkdfull.py fit").

Run XDKInterface.py from the tutorials folder (or fix some links).

Built With

Share this project: