Architecture of Parking Lots Management System for Drivers' Guidance

The goal of our project is to explore the already installed video surveillance cameras in the cities to provide an intelligent vision based approach for vacant parking lots detection and drivers' guidance. Our system should be able to treat huge number of parking stations in the town and to supply a real time information about the vacancies and their locations. The concept of distributed and extensible system is dominant in our case, so the need to the adoption of a Multi Agent implementation to ensure the interoperability, the autonomy, and the ability to take initiative and to interact with the external environment. In this paper we will present a general architecture of a Multi Agent Parking Lots Management system based on vision techniques dedicated to cover multiple parking stations at a city level. We will also demonstrate the importance of the elaboration of multi agent approach at this level to benefit from the cooperation and negotiation between agents in order to ameliorate the results of the already existent approaches.