Suggested Citation:
Moore, Jason, Mont Hubbard, A. L. Schwab, J. D. G. Kooijman (2010)

Accurate Measurement of Bicycle Parameters

Accurate measurements of a bicycle’s physical parameters are required for realistic dynamic simulations and analysis. The most basic models require the geometry, mass, mass location and mass distributions for the rigid bodies. More complex models require estimates of tire characteristics, human characteristics, friction, stiffness, damping, etc. In this paper we present the measurement of the minimal bicycle parameters required for the benchmark Whipple bicycle model presented in [7]. This model is composed of four rigid bodies, has ideal rolling and frictionless joints, and is laterally symmetric. A set of 25 parameters describes the geometry, mass, mass location and mass distribution of each of the rigid bodies. The experimental methods used to estimate the parameters described herein are based primarily on the work done in [3] but have been refined for improved accuracy and methodology. Koojiman’s work was preceded by [10] who measured a bicycle in a similar fashion and both [1] and [12] who used similar techniques with scooters.

We measured the physical characteristics of six different bicycles, two of which were set up in two different configurations. The six bicycles, chosen for both variety and convenience, are as follows: Batavus Browser, a Dutch style city bicycle measured with and without instrumentation as described in [6]; Batavus Stratos Deluxe, a Dutch style sporty city bicycle; Batavus Crescendo Deluxe a Dutch style city bicycle with a suspended fork; Gary Fisher Mountain Bike, a hardtail mountain bicycle; Bianchi Pista, a modern steel frame track racing bicycle; and Yellow Bicycle, a tripped down aluminum frame road bicycle measured in two configurations, the second with the fork rotated in the headtube 180 degrees for larger trail.

These eight different parameter sets can be used with, but are not limited to, the benchmark bicycle model. The accuracy of all the measurements are presented up through the eigenvalue prediction of the linear model. The accuracies are based on error propagation theory with correlations taken into account.