UC Irvine

The Two-fluid model has been used as a macroscopic method to evaluate urban network performance for decades. Different achievements have been made regarding to the use of the two-fluid model in city network and arterial street level. Data collected during peak hours is always used for parameter estimation, but how different congestion levels would affect the two-fluid model parameters has never been discussed. In this thesis, it is verified with ring road simulation data that under over-saturated or under-saturated traffic conditions, the two parameters, T_m and n, cannot be properly estimated. Beijing data is used in the second part to verify that GPS data from cellphone can be used for verification of the two-fluid model. Two data collection methods are used here: one is time-based, and the other distance-based. It is proved that with traffic data collected during peak hours, distance-based data is more accurate to estimate the two-fluid model parameters.