A car-following simulation model, CARSIM, with more realistic features to simulate not only normal traffic flow but also stop-and-go conditions on freeways, has been developed. The features of CARSIM are: (1) marginally safe spacings are provided for all vehicles, (2)start-up delays of vehicles are taken into account, (3) reaction times of drivers are randomly generated, (4) shorter reaction times are assigned at higher densities, and (5) dual behaviour of traffic in congested and non-congested conditions is taken into consideration in developing the car-following logic of this model. The validation of CARSIM has been performed at microscopic and macroscopic levels. At the microscopic level, the speed change patterns and trajectories from CARSIM were compared with those from field data; whereas at the macroscopic level, average speed, density, and volume computed in CARSIM were compared with the values from real world traffic conditions. The regression analysis of simulation results versus field data yielded R-squared values of 0.98 And higher, indicating that the results from CARSIM are very close to the values obtained from field data. One example of the application of CARSIM to study traffic-wave propagation is presented.
Samenvatting