Cellular signaling data driven simulation-based dynamic traffic assignment and its applications to a real-world road network

This paper presents a big data driven simulation-based dynamic traffic assignment (DTA) model by combining cellular signaling data, remote microwave sensors and license plate recognition (LPR) cameras data. First, the static origin destination (OD) matrix is estimated by using cellular signaling data. Second, remote microwave and LPR data are fused to calibrate parameters on the network supply side, including the link fundamental diagram calibration and link impedance function calibration. Third, a mesoscopic simulation model is developed by using DTALite, and the dynamic OD calibration is conducted by comparing link flows generated by simulation-based DTA and measured by remote microwave sensors. Finally, this paper takes the real-world road network of Hangzhou as an example to quantitatively assess impacts of large-scale implemented work zones before the 2016 G20 Summit and adverse weather conditions on the road network performance with the developed model. Results show the big data driven simulation-based DTA model developed in this paper has wide prospects of applications in the evaluation and optimization of comprehensive transportation management strategies.