A Testbed for Autonomous Racecar Control Strategies

The concept of a 'racing line' is an integral part of competitive racecar driving. While many drivers develop a sense for this optimal path around a racecourse intuitively, the physics governing basic vehicle dynamics around a path are fairly readily available. Naturally, this line is of great interest in racing game AI's and autonomous vehicle control, and a variety of strategies exist for developing a control algorithm that enables vehicles to closely follow this path. We modified a scale-model testbed consisting of a remote control vehicle, a networking system, and a computer backend enabling it to be controlled programmatically. The system can then be compared to and developed using a simplified model