Hello World, I would like to introduce a new interesting project to cip-labs. This project deals with a 1:10 RC car named Emmy. The goal of the project is to build up a car that is able to drive autonomously on the street without any human intervention. Furthermore, Emmy should be able to follow recognized objects and explore the environment.


It is not difficult to motivate this project. The latest developments in artificial intelligence allow some interesting use cases for the automotive industry, for instance: detection of moving objects and a lot of more relevant things. An ordinary car has a lot of controllers, sensors and so one. With Emmy the challenge are the limited hardware resources. Building a real-time system, that is able to control the behavior of a driving car (40km/h – 80km/h) is a very strong task and needs a lot of optimization.

Technical stuff

The car comes from Amazon. After the build up process of Emmy, we need a RaspberryPi (700Mhz, 512MB RAM) including a camera module to capture pictures and to process them. The RaspberryPi will be linked with the steering and engine of the car using a breakout board. To avoid unwanted collisions and track the road we are using a picture stream with 640×480 pixels up to 90 frames per seconds which should allow us to drive 50km/h or more. To process this data in real-time we plan to use an external computing source connected via Wifi stick. Below you can see a list of all necessary components:

  • Emmy – the car
  • RaspberryPi & RaspberryPi camera module
  • Wifi module to control the car via notebook
  • Status LEDs to check the confidence of cars estimation

Intelligent behavior

Building a RC car from the scratch is not a challenge at all. The main goal is to bring some intelligent algorithms to Emmy so that the car is able to drive smart. Some key features are:

  • Drive and remain on the street (straight, curve, crossing)
  • Collision control (using infra-red or grayscale)
  • Detect objects and follow them
  • Explore the environment without crashing
  • Map the environment to a digital map to revisit places

For some of these tasks, a few algorithms exist. The main purpose is to find an optimal algorithm that runs robust on the limited hardware.


I’m not alone. For the first time on cip-labs, this project runs with a cool project team. Gentlemen, a warm welcome to you!

Alexander Bresk – Initiator/Coding Artificial Intelligence

Sebastian Kuhz – Car Engineering

Felix Wolff - Electrical Engineering

Paul Wersch – Case Modding/Design of Emmy

To cut my huge enthusiasm about this project in a short sentence: I’m really excited and hope that we can achieve all goals with Emmy!

Leave a Reply

Your email address will not be published. Required fields are marked *


You may use these HTML tags and attributes: