Skip to Main content Skip to Navigation
Theses

Approche par invariance positive et les techniques de l'intelligence artificielle pour la régulation des carrefours signalisés

Abstract : Traffic control in a signalized intersection relates generally two distinct objectives: the thinning or reduction of congestion. In the first case, we avoid ending up in a situation of heavy traffic trying to adjust the durations of switching lights depending on the demand for attendance at the crossroads: it is an action a priori. In the second case, one is faced with a saturated traffic (congestion state). In this case, it will act retrospectively. In this work, we focus mainly on upstream work (action a priori) to avoid congestion by forcing queues to not exceed the level of traffic corresponding to the optimum operational lines. Specifically, modeled after the system, we propose a state feedback control based on the concept of positive invariance of sets and to achieve the objective. Two approaches are used: The first uses the Linear matrix inequalities (LMI). The second approach uses the concept of (AB)-invariance from the generalization of the theorem of Farkas. Then, we enrich both approaches by the technique of neural networks to estimate the inflow at the crossroads to ensure real-time feasibility of the proposed control. Finally, the results of this work are applied to a real intersection of the boulevard Anatole France to show their interest.
Document type :
Theses
Complete list of metadata

https://tel.archives-ouvertes.fr/tel-00720655
Contributor : Abes Star :  Contact
Submitted on : Wednesday, July 25, 2012 - 12:13:39 PM
Last modification on : Tuesday, October 1, 2019 - 3:37:12 PM
Long-term archiving on: : Friday, October 26, 2012 - 2:40:27 AM

File

These_MOTAWEJ_-_Fadi_UTBM.pdf
Version validated by the jury (STAR)

Identifiers

  • HAL Id : tel-00720655, version 1

Collections

Citation

Fadi Motawej. Approche par invariance positive et les techniques de l'intelligence artificielle pour la régulation des carrefours signalisés. Ordinateur et société [cs.CY]. Université de Technologie de Belfort-Montbeliard, 2012. Français. ⟨NNT : 2012BELF0180⟩. ⟨tel-00720655⟩

Share

Metrics

Record views

799

Files downloads

414