On human competitiveness of the evolved agent operating a scale model of a car

We present an approach for evolutionary design of the driving style of an agent, remotely operating a scale model of a car in a human competitive way. The agent perceives the environment from an overhead video camera and conveys its actions to the car via standard radio remote control transmitter. In order to cope with the video feed latency we propose an anticipatory modeling in which the agent considers its current actions based on the anticipated intrinsic (rather than currently available, outdated) state of the car and its surrounding. We formalized the notion of driving style by defining the key parameters, which describe it, and demonstrated the feasibility of applying genetic algorithms to evolve the optimal values of these parameters. The optimized driving style, employed by the agent, is human competitive in that it yields both faster and more consistent lap times than those of a human around a predefined circuit. Presented work can be viewed as a step towards the automated design of the control software of remotely operated vehicles capable to find an optimal solution to various tasks in a priori known environmental situations. Also, the results can be seen as a verification of the feasibility of developing a framework of adaptive racing games in which the human competes against a computerized opponent with matching capabilities, both operating physical, scale models of cars.