This paper considers a novel method of parametric identification of dynamic models, that is particularly appropriate for vehicle dynamics. The aim is to provide best possible estimates for parameters such as mass, stiffness and damping on the basis of experimental data. The particular issue addressed, is the robustness of the identification procedure to data corruption arising from unmodelled dynamic modes. The new method is based on a linear regression of impulse-momentum equations, obtained as the first integral of the ordinary differential equations of motion. Using randomised time intervals for the integration process, the new method is found to provide useful benefits. A simplified form of the method is also proposed. Both forms of the new method are applicable to both linear and non-linear dynamic models.
展开▼
