Analysis and application of two-fluid model for mixed traffic conditions

Two-fluid model quantifies traffic performance on a network by studying the interaction between moving and stopped vehicles in the traffic stream. The broad objective of this study is to investigate the presence and possible impact of measurement and modeling errors (non-ergodicity and endogeneity) on two-fluid model estimates under mixed traffic conditions. This objective is pursued based on travel time and running time data obtained from six-lane roads in Chennai city. To address non-ergodicity, a suitable temporal disaggregation scheme is proposed. An orthogonal regression model is applied that accounts for measurement errors in running times and trip times. To handle endogeneity, omitted variable bias correction is applied by adding two variables. The proposed non-ergodic model, correcting for both measurement error and endogeneity, provides a better fit and more accurate estimates than conventional model. These measurement and modeling errors also affect critical traffic flow parameters and level of service benchmarks.