A nonlinear frequency-dependent spring-mass model for estimating loading caused by rhythmic human jumping

Abstract

An empirical nonlinear, frequency-dependant, spring-mass system is conjectured for modelling human rhythmic jumping. This model is vital for correctly estimating human-structure dynamic interactions. An experimental study was employed to evaluate the leg mechanics and dynamic loading of a human jumper. Testing was performed over a large range of prescribed jumping frequencies. Subjects performed rhythmic jumps on a force plate and they were monitored by a motion capture system from which the displacement of the centre of mass was identified. Least squares system identification was utilised to determine the parameters of the spring-mass model for human rhythmic jumping. A nonlinear stiffness, rather than a conventional linear spring, is proposed to better capture the observed behaviour during periodic jumping. Force-displacement curves of each subject, during the contact phase of rhythmic jumping, were explored. These display an array of both classical Duffing’s type nonlinear softening and hardening spring stiffnesses over the range of jumping frequencies. The coefficients of the Duffing’s type model are observed to be highly sensitive to jumping frequency. A Poincaré section (phase-space) representation is used to visualise the jumping attractor’s topology. Thus, an experimental bifurcation analysis is performed suggesting the presence of both period doubling and fold bifurcations. These describe the transition from observed period-2 to period-1 jumping and coexisting low/high amplitude jumping behaviour. This study presents a framework for characterising the nonlinear loading of a human performing rhythmic jumping from direct measurements of force and displacement.