Machine Learning-Driven Analysis of Suspension Parameter Effects on Two-Wheeler and rider Vibrations on Class C Roads

Abstract

The dynamic behavior of a two-wheeler suspension system significantly affects ride comfort, stability, and human body vibrations, especially on uneven road surfaces. This study offers a mathematical modeling and simulation analysis of a two-wheeler with a seated rider navigating a Class C road profile, in accordance with ISO 8608 requirements. The model integrates essential suspension characteristics, such as tire stiffness (both front and rear), suspension spring stiffness (both front and rear), and damping coefficients (both front and rear), which were methodically altered to examine their effects on vehicle and rider dynamics. A Simulink-based multi-degree-of-freedom (MDOF) system was created to mimic vertical acceleration and displacement reactions across many body parts, including the head, upper torso, lower torso, viscera, and seat. A parameter sweep was performed across 729 distinct suspension configurations, and the resulting time-domain responses were analyzed to determine the maximum acceleration and displacement magnitudes for each body location.

Multiple regression-based modeling techniques were utilized to numerically describe the impact of suspension parameters on the observed biomechanical responses. The efficacy of many regression models was assessed to determine a precise predictive correlation among suspension stiffness, damping characteristics, and the resulting rider acceleration and displacement. The established models offer a computational framework for analyzing two-wheeler ride dynamics and serve as a basis for future research in suspension tuning, optimization, and human-centric ride comfort assessment.