Parameter identification of pedestrian's spring-mass-damper model by ground reaction force records through a particle filter approach

• Published in: Journal of sound and vibration 2017-12, Vol.411, p.409-421
• Main Authors: Wang, Haoqi, Chen, Jun, Brownjohn, James M.W.
• Format: Article
• Language: English
• Subjects: Biomechanical load factor; Biomechanics; Coefficients; Damping; Damping ratio; Dependence; Filters; Fourier analysis; Fourier series; Human-induced load; Identification methods; Load; Mass balance models; Mass-spring-damper systems; Mathematical models; Mechanical properties; Normal distribution; Parameter identification; Particle filter; Resonant frequencies; Spring-mass-damper model; Stiffness; System identification; Walking
• ISSN: 0022-460X; 1095-8568
• DOI: 10.1016/j.jsv.2017.09.020

The spring-mass-damper (SMD) model with a pair of internal biomechanical forces is the simplest model for a walking pedestrian to represent his/her mechanical properties, and thus can be used in human-structure-interaction analysis in the vertical direction. However, the values of SMD stiffness and damping, though very important, are typically taken as those measured from stationary people due to lack of a parameter identification methods for a walking pedestrian. This study adopts a step-by-step system identification approach known as particle filter to simultaneously identify the stiffness, damping coefficient, and coefficients of the SMD model's biomechanical forces by ground reaction force (GRF) records. After a brief introduction of the SMD model, the proposed identification approach is explained in detail, with a focus on the theory of particle filter and its integration with the SMD model. A numerical example is first provided to verify the feasibility of the proposed approach which is then applied to several experimental GRF records. Identification results demonstrate that natural frequency and the damping ratio of a walking pedestrian are not constant but have a dependence of mean value and distribution on pacing frequency. The mean value first-order coefficient of the biomechanical force, which is expressed by the Fourier series function, also has a linear relationship with pacing frequency. Higher order coefficients do not show a clear relationship with pacing frequency but follow a logarithmic normal distribution. •Particle filter approach is used to identify the parameters of a spring-mass-damper model for a walking pedestrian.•It is a parameter identification problem with unknown excitation.•Relations between human parameters and pacing frequency are given.