Solving route planning problems with experiences

In this paper, a scheme is outlined within which the autonomous vehicles in an intelligent transportation system exchange information about their environment. Benefiting from each other's experiences, the vehicles can locally adapt their behaviour in order to perform the desired flexible adjustments of the overall system behaviour. Based on the vehicles' shared knowledge, an exploration and planning strategy for selecting routes according to a set of given customer constraints is described. Meanwhile, sensor input from the environment is used to cooperatively learn the best routes and keep the knowledge about the world up-to-date.