Statistical analysis of floating-car data: an empirical study

We present the results of statistical analysis of the empirical floating-car data. Our investigations are based on analyzing the time series of four basic quantities namely velocity, velocity difference, spatial gap and the acceleration associated to some instrumented cars. We obtain the statistical characteristics, including the mean, variance and relative variance of these time series by taking direct time averages. We also try to identify the moving phases of the instrumented vehicle according to the statistical properties of its velocity time series. Moreover, by exploring the two-point joint probabilities, we propose a new approach for modelling vehicular dynamics based on the floating car data.