Modeling of Moguls on an Endurance Test Course

Abstract

This paper presents an approach to modeling discrete features on a U.S. Army endurance test course. The features in this case are 39 moguls with heights varying between 0.36 and 0.63 m built into a portion of the course. Because they are enumerable and localized, they do not lend themselves to traditional modeling techniques such as RMS, PSD, IRI, etc. We assume that the mogul has a well defined shape and we evaluate five different approaches to modeling this shape assuming that each mogul may be thought of as an ideal shape superimposed with stationary noise. The models that we evaluate are the Gaussian function, the Hanning window, the support vector machine, a two-Gaussian function and a two-Gaussian function with fixed width and amplitude ratios. We evaluate these models by computing the RMS error of each best fit and by evaluating the stationarity of the residue. We further evaluate each mogul model by running a HMMWV dynamics model over each and comparing several responses to those obtained from the profiled mogul. In our analysis we find that a model consisting of two Gaussian functions with related widths and amplitudes yields an unbiased estimate of a mogul and can be made to approximate any mogul by adjusting its width and amplitude.

Open PDF

Document Details

Document Type
Technical Report
Publication Date
Jun 01, 2007
Accession Number
ADA472154

Entities

People

  • David D. Gunter
  • Mark Brudnak
  • Wesley Bylsma

Organizations

  • Tank-automotive and Armaments Command

Tags

Communities of Interest

  • Energy and Power Technologies

DTIC Thesaurus Topics

  • Algorithms
  • Amplitude
  • Automotive Industry
  • Chassis
  • Curve Fitting
  • Deflection
  • Dynamic Response
  • Dynamics
  • Frequency
  • Frequency Domain
  • Kernel Functions
  • Measurement
  • North America
  • Roughness
  • Statistics
  • Supervised Machine Learning
  • Test And Evaluation

Readers

  • Approximation Theory.
  • Speech Processing/Speech Recognition.

Technology Areas

  • AI & ML
  • AI & ML - Bayesian Inference