ExoOpt - A framework for patient centered design optimization of lower limb exoskeletons

Powered exoskeletons have widespread future rehabilitation, medical and military applications. Current exoskeletons are challenging to use. They dynamically influence the human user such that he has to adapt his motions to the exoskeleton. This is due to the fact that the potential arising from an optimal integration of mechanics, electronics and control software has not been fully exploited during the design phase. In addition, interaction with the human user has not been considered to a full extent. We have developed an integrated design approach - implemented in the framework ExoOpt - that uses whole-body optimal control to simultaneously simulate a person and the exoskeleton he is wearing. Using this approach we have completed an optimal control study for a series of human and exoskeleton masses and inertias to identify the required kinematics and actuation of the exoskeleton during level-ground and slope walking. Our results can give important insights for the choice of actuators and the structural loads to be expected in the exoskeleton. They also show how much would be lost in terms of motion imitation capabilities if certain actuators such as the inversion/eversion direction of the ankle were replaced by a passive component. The presented approach is very flexible and can be applied to perform optimality studies to adjust geometry and mass distribution of exoskeletons, and can also be extended to other exoskeleton configurations.