Cost-effective vision systems for intelligent vehicles

The goal of this project is to develop cost-effective vision systems for a number of intelligent vehicle applications. We propose concurrent design of algorithms and hardware as a way of improving the cost and performance of vision systems. With this approach, we develop custom hardware rather than using off-the-shelf microprocessors. This paper describes three example systems which we have developed or are developing: car following, lane detection, and stereo vision. With the car following system, a driver pushes a button when he or she has a car in front which he/she would like to follow. The vehicle speed is controlled automatically using a vision system to keep a constant distance to the car in front. This system is described together with the lane sensing systems.