A Situational Awareness Approach to Intelligent Vehicle Agents

Abstract

As an increasing number of technological developments are made in the field of autonomous vehicles, the question of what intelligent system(s) will be placed around these vehicles both for the pursuit of individual goals and conformance to regulations as part of a wider collective of vehicles becomes pertinent, especially in the context of a mixed environment of autonomous and human controlled vehicles. The requirement to conform both to the law and with social conventions, in unpredictable circumstances, poses the problem of how to encode such knowledge. This paper adopts a Situational Awareness approach to agent knowledge, from low level perceptions, through to high level projection of future events, and explores a number of traffic scenarios where agents adopt different plans based on expected future states. A variant on such reactions is also presented, where the use of institutional governance frameworks is adopted to enforce certain behaviour, offering a ‘late binding’ mechanism for socially complex situations.