Analysing human walking behaviour using dynamic optimisation

Research in human walking can be dated back to many years ago, however a majority of the research has been empirical and few have considered explanatory or predictive assessment. Only in recent years have researchers considered modelling and simulating human walking motions. Even so, only selected portions of the human body were considered, and seldom did they look at modelling a full step cycle of walking. Walking, to date, remains as one of the more challenging problems to model. A mathematics model to simulate human walking motions and study the dynamics behind walking was formulated. The model is adjustable to accommodate different cases such as the single support phase and the double support phase of walking. Joint moment estimates were first calculated from the model and prescribed position coordinates of body segments using the method of inverse dynamics. The obtained values were used as initial joint torques which were required when using the method of dynamic optimisation to evaluate the model. Using dynamic optimisation, joint torque estimates can be improved and a better walking motion can be prescribed. Joint moments and forces were computed using the optimisation software MISER3, with the setup of appropriate objective functions and constraints. Experiment Main investigated whether the model formulated is able to produce normal walking motions and examined the optimal joint torques calculated to produce the resulting motions.