The autonomous car is the product of a joint project of Freie Universität and the Berlin Police Department. The result of the project is the “Spirit of Berlin” – a mini-van equipped with sensors and video cameras. It is possible for the driver to operate the car him/herself or, by pressing a button, to turn the operation of the car over to the computer. Utilizing special electronics, the computer can steer, brake, and accelerate the car, as well as turn all its components on or off.

Various sensors on the car, in combination with each other, can recognize pedestrians, cars, motorcycles, and other road users. This occurs partially with the aid of special laser scanners installed on the car. The scanners emit laser beams around the car that are reflected from objects in its proximity. The sensors measure the flight time of the reflecting laser beam. This shows where objects are located in the proximity of the car, as well as their distance from the car. Thus, the computer has an accurate map of the mobile and immobile objects on the road. The laser sensors’ range is 150 meters.

With the aid of its GPS-navigation system, which is quite a bit more precise than the usual GPS-systems, the autonomous car “Spirit of Berlin” can find its way through traffic better than any taxi driver. During a temporary loss of its satellite signal – e.g., due to high-rise buildings – the navigation system can precisely assess the position of the vehicle. To do this, data from gyroscopes and throttle sensors are processed. The computer can determine the position of the vehicle with a tolerance of error of one meter at the maximum.

It is crucial for the autonomous control that all data are processed fast and efficiently. This occurs with the aid of four computers donated by IBM. The IBM printed circuit boards receive all the data from the sensors and process the GPS-position data as well as the measurements from the laser scanner. In addition, the “Spirit of Berlin” has video cameras on the roof to aid in determining the position of the road markings and sidewalks. A special computer processes the video data. Finally, all the data from the sensors are combined, and a decision is made for the car’s next action.

The “Spirit of Berlin” was accepted as a participant for this year’s Urban Grand Challenge – the race for car robots. The race will take place on November 3 in California. Autonomous vehicles are only accepted for this race after passing a series of qualifying tests.

It will be a number of years before autonomous vehicles can be put on the market. Before that, driver assistance systems will be introduced gradually. Innovations such as automatic parking, adaptive cruise control, or collision warnings are already available – additional innovations will follow. This project of Freie Universität offers a preview of the future of mobility in urban traffic.

The autonomous car “Spirit of Berlin” was developed by a team around the computer scientist, Raúl Rojas, a professor for artificial intelligence at Freie Universität Berlin since 1997. The native of Mexico is known, among other things, for the international success of his soccer-playing robots, “FU-Fighters,” that have won the world championship for robots, RoboCup, twice. In 2005 Professor Rojas received the Wolfgang-von-Kempelen-Prize for History of Computer Science, awarded for the first time by the Austrian Society for History of Computer Science (OeGIG), the Austrian Computer Society (OCG), and the Austrian Federal Department of Transportation, Innovation, and Technology. His main field of research is artificial neural networks.

Additional sponsors are still needed for the “Spirit of Berlin” project.

The project is described in more detail online:

http://robotics.mi.fu-berlin.de

Additional information is available from:

Prof. Dr. Raúl Rojas, Institute of Computer Science of Freie Universität Berlin, Telephone: 030 / 838-75130, Email: rojas@inf.fu-berlin.de