Parameter Identification of Pedestrian Models on Flexible Bridges: An MSDE Regression Model and a Comparative Applicability Framework

Pedestrian bridges with low damping, slender, and lightweight characteristics exhibit pronounced human–bridge interaction (HBI) effects. For the coupled human–bridge system, both pedestrian dynamic properties and pedestrian load are affected by bridge vibration. Although different models were proposed by researchers for pedestrian dynamic properties and pedestrian loads, most of them were based on tests on rigid ground or vibrating bridges with high fundamental natural frequencies, which do not reflect the HBI effect. The applicability of different models may vary with specific scenarios. To establish a more rational pedestrian model for footbridges prone to HBI, in this study, human-induced vibration tests were conducted on a flexible and low-damping simply supported footbridge. The theoretical transfer function of the an external harmonic force moving with the mass-spring-damper (MSDE) model, composed of the pedestrian mass spring damper and an external harmonic force, was derived for the case of a person walking across a bridge. The genetic algorithm combined with SIMULINK (version 10.6, part of MATLAB R2022b) was used to perform a closed-loop study for determining pedestrian dynamic parameters and harmonic forces. By developing a novel MSDE regression model and a comparative applicability framework, the Human–Bridge Coordination Index was proposed to serve as an effective prescreening indicator for model applicability. The proposed MSDE regression model may be considered for preliminary analysis in the absence of prior experience. This work provides a fundamental approach to managing parameter variability in the vibration serviceability assessment of footbridges.