Non-Invasive Identification of Vehicle Suspension Parameters: A Methodology Based on Synthetic Data Analysis

Abstract

In this study, we introduce an innovative approach for the identification of vehicle suspension parameters, employing a methodology that utilizes synthetic and experimental data for non-invasive analysis. Central to our approach is the application of a basic local optimization algorithm, chosen to establish a baseline for parameter identification in increasingly complex vehicle models, ranging from quarter-vehicle to half-vehicle (bicycle) models. This methodology enables the accurate simulation of the vehicle dynamics and the identification of suspension parameters under various conditions, including road perturbations such as speed bumps and curbs, as well as in the presence of noise. A significant aspect of our work is the ability to process real-world data, making it applicable in practical scenarios where data are obtained from onboard sensor equipment. The methodology was developed in MatLab, ensuring portability across platforms that support this software. Furthermore, the study explores the application of this methodology as a tool for denoising, enhancing its utility in real-world data analysis and predictive maintenance. The findings of this research provide valuable insights for vehicle suspension design, offering a cost-effective and efficient solution for dynamic parameter identification without the need for physical disassembly.