Developing compact autoregressive characterisations of terrain topology profiles

Abstract

Terrain is the principal source of vertical excitation into the vehicle system, but it is impractical to simulate long distances of every terrain type. A terrain characterisation method is developed for grouping terrain into sets with similar characteristics. Profiles are considered realisations of a stochastic process, described by an autoregressive (AR) model and characterised by the AR coefficients and a single cumulative probability function for the residual process. A set of basis vectors is developed onto which the inverse cumulative probability function (ICPF) of the residual process is mapped. The ICPF is estimated with a small number of basis vectors. The AR and ICPF coefficients of the residual process are jointly dependent on the underlying terrain profiles, so additional basis vectors that map these dependent coefficients onto a reduced set of coefficients are developed. It is shown that six coefficients are required to characterise US highways.