Berlin, Germany – Researchers at a university in Berlin have developed software which allows a driver to steer a car with his eyes.
A team from Freie Universität’s Artificial Intelligence Group demonstrated the system, installed in a Dodge Grand Caravan named the Spirit of Berlin. The software collects information on the driver’s eye movements and converts it into control signals for the steering wheel. The speed is controlled separately and not included in eyeDriver.
A converted bicycle helmet equipped with two cameras and an infrared LED looks simultaneously in the same direction as the person wearing the helmet (the scene camera), and at the wearer’s eye (eye camera). The software captures the position of the pupil in the eye camera and calculates the position in the scene camera, transmitting them to the onboard computer in the vehicle and using the information to control the steering wheel.
The system includes “free ride” and “routing” mode. In free ride mode, the viewing positions are linked directly with the steering wheel motor, and used to calculate the desired position of the steering wheel. The further the driver looks to the left or right, the further the steering wheel is turned in that direction. If it is not possible to detect the direction, such as if the driver’s eyes are closed, the vehicle automatically brakes.
In routing mode, the vehicle steers autonomously most of the time, and only asks the driver to select a route if it comes to an intersection or fork in the road. If the driver looks long enough in one direction, the eyeDriver software confirms the choice.
The Spirit of Berlin made it to the semi-finals in the 2007 DARPA Urban Challenge in California for autonomously-driven vehicles. In 2009, researchers also developed the iDriver, which made it possible to steer the research car using an iPhone. The various technologies are part of a modular system for autonomous or semi-autonomous cars aimed at developing collision-free, efficient and environmentally-friendly mobility.