In this project the PhD candidate will apply data analytics on the Bluetooth trajectories and traffic states to empirically estimate the assignment matrix for the network, generating new knowledge and factors affecting driver route choice behaviour and benchmark analytical route choice models.
Furthermore, an understanding of traveller’s behaviour helps develop policies for infrastructure expansion and development including the route choice under the influence of an incident or understand travel patterns for major events such as the Olympics.
Traffic assignment uses route choice models and user/system equilibrium principle to assign the demand on the network. In literature, route choice models are calibrated from limited surveys and equilibrium is achieved using simulation.
However, attempts to exploit large-scale, ‘real’ traffic monitoring data to calibrate route choice models and/or empirically model traffic assignment are limited. In this research we will apply data analytics on vehicle trajectories and traffic states to empirically estimate the assignment matrix for the network, generating new knowledge and factors affecting driver route choice behaviour and benchmark analytical route choice models.
This provides a paradigm shift in real time simulation where the modelled assignment is based on the evidence from the observations and not from computationally expensive equilibrium. The calibrated models provide an opportunity for the traffic modelers to forecast the traffic conditions and travel patterns, thereby developing long and short-term traffic control strategies.
Objective 1: Evaluation of the various route choice models in practice
Objective 2: A data-driven process of extracting path utilities and developing calibrating route choice models for traffic assignment
Objective 3: Testing of the fully-functioning calibrated route choice model in traffic assignment
More from iMOVE