A Fuzzy Based Driving Architecture for Non-player Characters in a Car Racing Game

Videogame-based competitions are receiving a lot of interest in the last few years, as they provide an ideal framework to apply soft computing techniques. One the most popular competitions is the Simulated Car Racing Competition and thanks to the realism implemented by recent car simulators, it provides a test framework for the application of autonomous driving techniques. This work presents the design and implementation of a car controller able to deal with competitive racing situations. The full driving architecture is composed by six simple modules being each one responsible for a basic aspect of car driving. The modules control gear shifting, steer movements and pedals positions by using of simple functions meanwhile the speed control is managed by a simple fuzzy system. Additionally, two modules are in charge of: 1) adapting the driving behavior when other cars are present; and 2) implementing a basic ’inter-lap’ learning mechanism in order to remember key track segments and adapt speeds in future laps. The assessment of the controller was done in two ways. Firstly, running without adversaries, where our controller allowed to race the longest distance and secondly, participating in the 2009 Simulated Car Racing Competition where it won the competition.