On cognitive management and non-causal reasoning for enabling highly automated driving

Abstract

Highly Automated Vehicles (HAVs) have become a trend, and also a hotspot of research in recent years, aiming to support, or even to replace, human drivers. Their goal is mainly to strengthen the driver's sensing ability and to reduce the control efforts of the vehicle itself. Moreover, on-board communications equipment helps vehicles to have a model of their complex driving environment that includes the presence (meta-knowledge) of other entities sharing the same driving scene. Therefore, cognitive decisions should be taken in an automated manner, being able to operate HAVs each time in the best available Level of Autonomy (LoA), by responding quickly not only to causal reasoning effects, which depend on present and past inputs from the external driving environment, but also to non-causal reasoning situations, which depend on future states associated with the external driving scene. The present study aims to tackle exactly this challenge by introducing an on-board cognitive decision-making functionality, which operates on the basis of collecting information from various sources, intelligently processing it, integrating knowledge and experience and, finally, selecting the optimal LoA.