Accurate travel time estimations are needed in many intelligent transportation systems (ITS) applications. This work describes a prediction and smoothing algorithm, based on the calculation of the power spectrum of the daily courses from floating car data (FCD). The so called Lomb Periodogram delivers the frequencies, which are needed to describe the daily course with a simple trigonometric function, the function parameters are approximated with the help of singular value decomposition fit (SVD-fit).
Samenvatting