Traffic Planning and Coordination of Automated Vehicles
Date
Type
Examensarbete på kandidatnivå
Programme
Model builders
Journal Title
Journal ISSN
Volume Title
Publisher
Abstract
Systems for autonomous vehicles can increase productivity within the transport
industry. Automated driving systems will liberate people from the need to drive,
allowing the driver to be engaged with other vocational activities. Moreover, they
will allow people and products that cannot drive to travel without the need to
hire driver services. In order for such a system to work, a solution that can
guarantee efficiency and safety is required. A proposal for a coherent system for
planning and coordinating automated vehicles is made. This is accomplished by
using multi-agent path finding to plan optimal routes for all vehicles in a closedoff
road system, and coordinating their journeys in a safe manner with the help
of a collaborative driving management system. The designed system consists of a
continuous conflict-based search algorithm that can plan collision-free paths for up
to 100 vehicles simultaneously. The virtual traffic light proposed uses an agreement
system where it collaborates with agents to determine paths that should be given
the green light. By also designing a service allowing vehicles to connect to the
virtual traffic light server corresponding to an upcoming intersection, a safe and
optimal solution is achieved. The proposed system can act as a holistic pilot
system for autonomous vehicles in closed-off environments such as airports, ports,
and mines.
Description
Keywords
Autonomous Driving, Virtual Traffic Lights, Conflict Based Search, Multi-Agent Path finding