A Preliminary Investigation of an Autonomous Vehicle Validation Infrastructure for Smart Cities

Abstract

The research and development of autonomous vehicle has entered the era of commercialization. While the vehicle self-driving technology has been growing rapidly, the validation for autonomous vehicle in terms of driving model, human factor model and traffic model is still maturing. Most of previous infrastructures are mainly focused on validation of those three models separately resorting either on real driving test at physical infrastructure or software simulation in virtualized infrastructure. However, neither the real driving test can cover all possible scenarios of autonomous driving and human factors, nor the virtualized software simulation can generate a feasible model for practical on/off-road driving. Furthermore, future autonomous transport in smart cities requires comprehensive validation. In order for autonomous vehicles to meet the autonomous transport in such complex traffic environment, an integrated testing and simulation infrastructure has been built targeting the systematic validation for autonomous vehicles: the Multi-User Environment for Autonomous Vehicle Innovation (MUEAVI). A preliminary investigation of a new autonomous vehicle validation infrastructure that can serve a multitude of research projects for smart city is presented.