Laboratoire de Génie Informatique et d’Automatique de l’Artois


Ph.D. student
(Position when this person departed from the LGI2A)
Member of the research axes:
No publication meets the provided criteria.

Author of the Ph.D. thesis "Commande non-entière des systèmes de commande, Développement et application pour les modèles de flux de trafic routier"

2006 - 2009

This thesis presents research carried out to several elements of the macroscopic traffic flow as the model, the control and the simulation of his control system. The main aims of the realized studies consist to keep the circulation on the high-ways fluid. That means that we must to assure some quality of the process regarding the stability of this process. More over to offer best performances and quality of the traffic services for the users on the ways networks.

In our study we use the analytical solution method of the dynamic equation presenting the LWR traffic flow model process, for which we look to obtain transfer function. Our objective is to obtain a conform result to a toll plaza. Furthermore we look to make a choice of appropriate control algorithm to satisfy the traffic network and users’ needs. The traffic flow management needs results from the increasingly of the flows. As consequence of this we can obtain saturation in some places in the road network wildly known as a traffic jam usually in the rush hours, by reason of accident or repairs works. All this provoke a delay of the transportation flow and important environmental after-effect. Therefore it’s very important to assure the fluidity of the traffic using control strategies which will cancel, reduce or delay the traffic jam appearances. Because of all the reasons above, we have proposed a system with non-integer order control algorithm for maintain the traffic fluid by the control of the pikes in the toll plaza. The control variable is the upstream density which will influence on the downstream one. After the analytical solution of the toll plaza model we obtain a delay function which presents the plant in our distributed parameter system. For this system we apply a Smith prediction non-integer control algorithm and moreover we ameliorate this system with a Dead time non-integer order compensator.