Qualitative Change of Car-Following Behavior Observed in Real Traffic

Summary

Microscopic dynamical properties of traffic flow are studied from the aspects of vehicle behavior and driver operation. We studied the nature of fluctuations around the critical region in real traffic by analyzing a time series of variations of velocity obtained from single-vehicle data measurement. We found that the probability density function calculated from the time series of velocity variations is transformed, while a Gaussian distribution transitions into a stable symmetrical Lévy distribution. The power-law tail in the Lévy distribution indicated that the time series of velocity variation exhibits critical fluctuations. The power-law tail in the probability density function suggests that dynamical processes of vehicular traffic are related to a time-discrete stochastic process driven by random amplification with additive external noise. In contrast, the empirical data of deceleration in a car-following situation obtained from the driving simulator experiment indicated a large dispersion of perceptual quantities of a driver when operating the brake pedal. The result suggests that the algorithm for operating the brake pedal is closely related to the random amplification in the discrete stochastic process.