Agent Based Traffic
Simulator for Autonomous
Vehicle
BY:
Kamalika Saha
Ms. Sonia Rathee
Deptt. of CSE, Maharaja Surajmal
Institute of Technology, New Delhi,
India
Introduction
•
The project aims to provide a simulation for traffic
management at intersections in case of autonomous
vehicles which can further be implemented at a citywide scale. The advent of driverless cars provides
optimizing traffic in ways which were not possible
before.
•
The project has been made to substitute the
conventional means of controlling traffic particularly,
the traffic signals.
Objective
•
This project aims to create a scalable, safe and efficient
multi-agent framework for managing autonomous
vehicles at intersection.
•
In the future with computers “behind the wheels”, it will
make travel easier, safer and much more efficient.
What’s Wrong With The
Conventional Traffic System?
•
Traditional traffic systems & stop signs are very inefficient.
•
Vehicles traversing intersections using these mechanisms
experience large delays.
•
The intersections can only manage a limited traffic
capacity – much less than that of the roads that feed into
them.
•
The time that vehicles spend in case of traffic systems
results in unnecessary wastage of fuel powering the
vehicles.
Autonomous Intersection
management
•
Autonomous Intersection Management (AIM) aims to provide
an alternative to existing traffic signal optimization
schemes.
•
In this approach, we use a program called the driver agent
located in each vehicle that communicates with the server
present at the intersection points and other agents in a
multi-agent environment.
•
It also uses an intersection manager which is assumed to be
a server present at each intersection.
The Reservation Idea….
•
Any intersection manager present at the intersection
maintains a reservation table for the list of past requests
made by any autonomous vehicle.
•
If there are no conflicts, the intersection manager issues a
reservation. It becomes the vehicle’s responsibility to
arrive at, and travel through, the intersection as specified.
•
In the case of a conflict, the intersection manager suggests
an alternate later reservation.
•
The car may only enter the intersection once it has
successfully obtained a reservation.
Implementation
•
The study of algorithms used in project implementation
such as Intersection Manager Algorithm, FCFS (First Come
First Serve), A* Search.
•
Development of GUI for a simulator using Java-Swing.
•
Implementation of the above described algorithms to
create a stable simulation.
Demo Screen Shots
Limitations

If the traffic at the intersection is increased beyond
certain limits then the simulator might not respond or is
not able to support intersection policies simulation .

Lane Changing Algorithm is not implemented in our
project.

Mechanical failures are not taken into account.
Conclusion
•
We, therefore, conclude that the project to create a
simulation for Agent Based Traffic Simulator for
autonomous vehicles has been implemented successfully,
however, there is a wide scope to implement the project
on a real time basis and implementing the lane changing
algorithm successfully
References

Approximately Orchestrated Routing and Transportation
Analyser: Large Scale Traffic Simulation for Autonomous
Vehicles.

Vehicle to Infrastructure based Safe Trajectory: Planning
for Autonomous Intersection Management.

Auction-based autonomous intersection management.