Methodology to convert a transportation planning origin-destination matrix into a microscopic traffic simulation origin-destination matrix

Abstract

Microscopic traffic simulation (MTS) is a cost-effective approach for the evaluation of traffic conditions in urban highways networks. In MTS models, traffic demand that is entered into the network is specified by one or several Origin-Destination (O-D) matrices. A major challenge in the application of MTS for large urban networks is the specification of the O-D matrices. This dissertation proposes a methodology that may be used to transform a O-D matrix from a transportation planning model (which is based on Traffic Analysis Zone, or TAZ, and is readily available) into a O-D matrix for a MTS model (which is traffic zone based, requires a higher resolution and larger matrix size).

The main concepts of the methodology are the discrete choice approach, followed by path analysis in a geographic information system. The discrete choice model predicts a driver's route choice preference based on the socioeconomic and trip characteristics of the driver. The path analysis matches the driver's route choice preference to an access point along a main street at the boundary of the TAZ. By applying the concept to every parcel in a TAZ, the volume and percentage splits of all the access points at the boundary of the TAZ is then estimated, from which the limited number of access points are selected to form the traffic zones and zone connectors for the MTS model. Once the traffic zones of all the TAZs have been identified, the trips generated from and attracted to each TAZ are proportioned among the traffic zones. The gravity model is applied to perform trip distribution analysis to produce the O-D matrix for the MTS model.

The proposed methodology has been developed based on a primarily residential TAZ within the City of El Paso, Texas. The morning peak hour home based trips made by drivers in the residential parcels in this TAZ have been predicted. The volume splits between all the access points have been validated against field traffic counts. To methodology was then applied to a network with five TAZs to demonstrate its use to construct a O-D matrix for MTS of a relatively larger network.

This research contributes to the application of MTS models to large urban networks by adapting the O-D matrix from transportation planning models, saving the analyst time to construct a O-D from scratch. The methodology is flexible enough to be transferred to other metropolitan areas and also able to forecast O-D matrices for future years.