Evolving a rule system controller for automatic driving in a car racing competition

The techniques and the technologies supporting automatic vehicle guidance are important issues. Automobile manufacturers view automatic driving as a very interesting product with motivating key features which allow improvement of the car safety, reduction in emission or fuel consumption or optimization of driver comfort during long journeys. Car racing is an active research field where new advances in aerodynamics, consumption and engine power are critical each season. Our proposal is to research how evolutionary computation techniques can help in this field. For this work we have designed an automatic controller that learns rules with a genetic algorithm. This paper is a report of the results obtained by this controller during the car racing competition held in Hong Kong during the IEEE World Congress on Computational Intelligence (WCCI 2008).