Characterizing 2-D Topographic Mappings of Roads

Abstract

Load data representing severe customer usage is needed throughout a chassis development program; the majority of these chassis loads originate with the excitation from the road. These chassis loads are increasingly derived from vehicle simulations, however simulating a vehicle traversing long roads is impractical and a method to produce short roads with given characteristics must be developed. The first step is to consider the road to be a realization of an underlying stochastic process. There are many methods currently available to characterize roads when they are assumed to be homogeneous. The issue of non-stationarity that arises when a vehicle traverses a homogenous road at a varying speed has also been discussed. This work develops of method of characterizing non-stationary road profile data using a pure autoregressive process. The model is developed utilizing the sample autocorrelation and partial autocorrelation functions. The adequacy of the model is evaluated through statistical diagnostic checks performed on synthetic road data generated by the autoregressive model parameters. Use of these parameters to classify roads is also discussed as possible future work. Any synthetic road realized from a given class of model parameters will represent all roads in that set.