A Simulation Framework of Remote Drive-Directing over Local Autonomous Vehicles for Experimentation

초록

This paper describes the simulation framework of a commanding system for a group of autonomous vehicles circulating in the same driving zone. The commanding system arbitrates their driving routes to ensure efficient sharing of the roads within the zone. The framework contains the essential entities to be considered in building such a commanding system, which does not currently exist. The commanding system directs the driving of each autonomous vehicle remotely, providing a functionality that has not yet been properly addressed in academia or industry. Accurate tracking of the vehicle locations is essential to commanding the autonomous vehicles in the zone. Environmental limitations, however, restrain frequent updating of the vehicles’ locations with real data coming from them. In this work, we examine the validity of calculating vehicles’ velocities in order to maintain the accuracy of their locations within some level, by comparing simulated velocities to emulated velocities through a simulator. In calculating the velocity, pseudo-parameters are used in the simulation, which are derived from parameters generated in the low-level driving computation about which we already have knowledge and experience.