000195975 001__ 195975
000195975 005__ 20190213063710.0
000195975 037__ $$aPOST_TALK
000195975 245__ $$aIntegrated Transportation and Urban Infrastructure Systems: Planning and Design
000195975 269__ $$a2012
000195975 260__ $$c2012
000195975 336__ $$aTalks
000195975 520__ $$aPlanning and design of urban infrastructure systems in general and transportation systems in particular, require a multidisciplinary effort that is heavily embedded in a system-of-systems perspective. The envisioned research direction is a bottom-up approach that is focused on understanding individual human-agents performing activities in urban systems and interacting with each other as well as with the infrastructure. The dynamics of the urban system is thus an emergent phenomenon out of the aggregation of individual behaviours. Such an approach can go beyond extending trends and can extensively address what-if questions as a result of policy and design changes. In particular, this talk will demonstrate how such an approach is useful in planning and design at a large-scale (i.e. urban region) and at a micro-scale (i.e. mobility hubs and public spaces). This seminar presents the breadth of urban systems research that the presenter, Bilal Farooq, has been engaged in with a focus on two major applications. At urban region level, a novel simulation-based population synthesis approach will be presented. Unlike conventional fitting-based approaches that formulate an optimization problem, here the aim is to directly draw from the underlying joint distribution. To do so the partial views are utilized that are available in the form of data samples and control totals. Using Swiss census, a comparative analysis with existing approaches will be discussed and the superiority of this novel method will be highlighted. At a micro-scale, a model for pedestrian activity-episode sequence will be presented. The approach relies on network evaluation data from WiFi access points. A Bayesian approach will be presented; it utilizes the prior on activities and travel times; and develops a measurement model based on WiFi triangulation in order to predict the likelihood of activity-episode sequences. A case study on EPFL campus will be illustrated.
000195975 700__ $$g207159$$aFarooq, Bilal$$0245136
000195975 7112_ $$dDecember 06, 2012$$cÉcole Polytechnique de Montréal, Montréal, Canada$$aInvited speaker
000195975 909C0 $$xU11418$$0252123$$pTRANSP-OR
000195975 909CO $$ppresentation$$pENAC$$ooai:infoscience.tind.io:195975
000195975 937__ $$aEPFL-TALK-195975
000195975 970__ $$aTALK-BF_PolyDec2012/TRANSP-OR
000195975 973__ $$aEPFL
000195975 980__ $$aTALK