Memory effects and scaling properties of traffic flows

Department of Physics, The Norwegian University of Science and Technology N-7491 Trondheim, Norway

Corresponding authors: Bo-Sture.Skagerstam@whys.ntnu.no Alex.Hansen@phys.ntnu.no

Received: 21 April 2005
Accepted: 28 September 2005

Abstract

Traffic flows are studied in terms of their local noise of sound, which is an easily accessible experimental quantity. The sound noise data is studied making use of scaling properties of discrete wavelet transforms and Hurst exponents are extracted. The scaling behavior is used to characterize the traffic flows in terms of scaling properties of the memory function in the framework of Mori-Lee stochastic differential equations. The results obtained provides for a new theoretical as well as experimental framework to characterize the large-time behavior of traffic flows. The present paper outlines the procedure by making use of data from one-lane computer simulations as well as sound data of a real traffic flow. The experimental real traffic flow data we have obtained, using standard sound measurement techniques, is compatible with the presence of conventional diffusion at small time scales as well as noise at large time scales.