Multi-agent Behavior in a GIS Traffic Simulation 

Presented by: David Nikaido & Tiffany HuntFaculty Mentor: Dr. Christine Drennon

Introduction

Literature Review

Agent

Enviroment

Future Work

Timeline

Conclusion

Modeling Traffic in a Multi‐Agent System with Geographic Information Systems (GIS)Trinity Summer REU 2009

Goals: Simulate traffic in San Antonio

◦ Agents

◦ Enviroment

Our goals are to build off of last years traffic simulation model in two ways:

◦ Create a citywide map structure

◦ Implement driving agent behavior

“A behavioral multi-agent model for road traffic simulation.” Doniec, Mandiau, Piechowiak and Espie.

Three types of behavior

◦ Normative

◦ Opportunistic

◦ Anticipatory

“A collaborative driving system based on multiagent modeling and simulations.” Halle’, Chaib-draa

Traffic Control LayerTraffic Control Layer

Management LayerManagement Layer

Guidance LayerGuidance LayerInputInput outputoutput

Graphical Information Systems(GIS)◦ Spatial◦ Data◦ Pretty

Types of agents:◦ Road agents◦ Traffic lights◦ Signs◦ Sensors

Path finding algorithm Shortest path Alternate path

Autonomous driver agents (DARPA) User functionality:

◦ Change traffic rules◦ Change driver behavior◦ Measure traffic flow◦ Change road structure

Week 3 - 4:◦ Translate last year’s code to Scala

Week 5 - 7:◦ Implement agent behavior◦ GUI

Week 8 – 9:◦ Optimize code and future work

Can be useful for testing traffic infrastructure

FUN!! Can be used to measure real world data