Minimal Statistical Equivalent for Virtual Public Road

Abstract

A minimal statistical equivalent for virtual public roads is developed for durability simulation. The road is a primary excitation source for vehicle loads and fatigue, but measured public roads are long, spatially uneven, and costly to use directly in repeated simulation workflows. The central question is therefore how far measured road data can be reduced while still remaining statistically equivalent to the original road environment for the intended vehicle response application. Statistical equivalence is interpreted as more than agreement in visual road shape or in a single roughness descriptor. It requires preservation of the road statistics that govern excitation relevant to response, including spectral content, nonstationary roughness variation, localized transient events, and lateral surface structure. Two modelling routes are compared. The Nonstationary Laplace route, denoted NSL, represents nonstationarity implicitly through a compact stochastic description of road severity. The Wavelet Multi Component Reconstruction route, denoted WMCR, separates and reconstructs topography, background roughness, transients, and two dimensional surface content more explicitly. The methods are evaluated through road statistics, reduced quarter car response, and nonlinear full vehicle simulation in Adams Car. The results show that compact statistical descriptors are useful for efficient virtual road generation, while localized transient structure must be treated carefully when fatigue relevant response is the target. A minimal statistical equivalent for virtual public roads should therefore be defined jointly by road statistics and vehicle response equivalence, rather than by compression, appearance, or average roughness alone.