Using ego-centered affordances in multi-agent traffic simulation.

Feirouz Ksontini (1) , Zahia Guessoum (2) , René Mandiau (3) , Stéphane Espié (1)

1 IFSTTAR - Institut Français des Sciences et Technologies des Transports, de l'Aménagement et des Réseaux
2 SMA - Systèmes Multi-Agents
3 LAMIH - Laboratoire d'Automatique, de Mécanique et d'Informatique industrielles et Humaines - UMR 8201

To improve the validity of traffic simulations in urban and suburban areas, we propose to consider the driving context and the driver behavior in terms of space occupation. We endow agent driver with an ego-centered representation of the environment. This representation permits the agent to take a decision in terms of space occupation. Our agent driver model is based on the concept of affordances - the ways in which an agent can interact with its environment. First, we use the concept of affordances to identify the possible actions, in terms of space occupation, afforded by the environment. Second, we use an ego-centered representation of the situation around the agent, composed by the identified affordances. The proposed driver model was implemented with ArchiSim and the experiments show that this model makes traffic more fluid.

Mots clés: "Affordances"; "Multi-agent simulation"; "Traffic simulation"; "Virtual lanes"

Domaines: Informatique [cs]

Feirouz Ksontini, Zahia Guessoum, René Mandiau, Stéphane Espié. Using ego-centered affordances in multi-agent traffic simulation.. 12th International Conference on Autonomous Agents and Multiagent systems, AAMAS 2013, May 2013, Saint Paul, Minnesota, United States. pp.151-158. ⟨hal-01217287⟩