Friday Seminar - Sebastien Blandin on Modeling, Estimation and Control of Distributed Parameter Systems

Istanbul highway traffic

This week’s Friday Seminar has Sebastien Blandin, PhD., Research Scientist, IBM Singapore Research Collaboratory, presenting “Modeling, Estimation and Control of Distributed Parameter Systems: Application to Transportation Networks.”

Scalar conservation laws expressing the evolution of traffic density have been extensively used for macroscopic traffic modeling and forward simulation using loop detector measurements. The mobile sensing paradigm allowing increased understanding of traffic flow motivates the need for novel partial differential equation results, of practical interest for real-time traffic operations. The research contributions of this work are centered on traffic modeling capabilities, data assimilation performances, and Lyapunov stabilizability guarantees. We propose a new 2X2 phase transition model of traffic flow able to account for classical traffic phenomena as well as heterogeneous driving behaviors. In the context of traffic information systems and filtering algorithms, we present one of the first numerical and analytical descriptions of the consequences of the emergence and propagation of shockwaves in traffic flow on estimates accuracy. Finally, we propose a Lyapunov stability result for the entropy solution to the transport equation using boundary actuation. The results presented in this work are instantiated on the Mobile Millennium system providing real-time traffic estimates in Northern California.

The seminar will take place on Friday, January 27, from 4:00-5:00 p.m. in 534 Davis Hall.