Modelisation of a Human-Exoskeleton Interaction for Cerebral Palsy

This paper presents a method to model a human-exoskeleton interaction for patients suffering from spastic cerebral palsy. We base our work on the Clinical Gait Analysis performed on two 9-year-old twin sisters. The first sister has a pathology called spastic cerebral palsy, while the second sister is non-pathological without any impairment. This paper aims at determining the proportion of the walking efforts that can be supported by an exoskeleton in order to allow a pathological child gait to converge toward a non-pathological one. Based on experimental data, a model of the pathological gait reconstructed using an optimal estimation. The model relies on mechanical differential equations of motion. The interaction between the human and the exoskeleton is then modelled using optimal control, while modelling ground contacts. Results show that the human produced joint torques are within the possible torque range of the child with cerebral palsy, which justifies the use of an exoskeleton to correct a pathological gait. The code for running the simulations is provided in open source.