Examining the Stationarity and Ergodicity of Traffic Flow on Highway Bridges

Abstract

According to the information of vehicles collected by weigh-in-motion (WIM) systems on 28 lanes of 6 highway bridges, stochastic processes of the type of traffic flow on each lane are generated respectively. Time means and time-correlation functions were calculated accordingly based on stochastic process theory. The stationarity and ergodicity of the composed stochastic processes of the vehicle type of traffic flow were then examined. The statistical characterization of samples collected by WIM systems on actual highways will approach the total value if the samples are sufficiently plenitudinous. Moreover, the statistical characterization of the collected vehicle information will exhibit the characters of the entire specimen. Such results can be used as guides to model random vehicle load further.