DeepTonomous

Inspiration

DeepTonomous is a Self-Driving RC Car that uses the latest Object Detection Algorithms to navigate autonomously

What it does

It follows lane markings and avoids obstacles by perceiving its environment

How we built it

We attached sensors on the Arduino board and 3D printed the chassis. The network was trained on the Faster R-CNN Object Detector and the ResNet101 backbone architecture.

Challenges we ran into

We had problems to print the chassis with a 3D printer because it was too big for it. We decided to split it into four parts.

Accomplishments that we're proud of

Every team member participated to the final result to build an Autonomous Car.

What we learned

We learned how to build a car with multiple sensors and how to train neural networks to detect obstacles on the road (e.g. cars, pedestrians and cyclists)