ExoOpt - a tool for evaluating exoskeleton designs using model-based optimization

We present a design study for a lower-limb exoskeleton using model-based optimization techniques. A physically realistic combined human-exoskeleton model is fitted to data of regular human walking motions in different conditions using a least squares optimal control formulation. This helps to determine which torques would be required to make the human-exoskeleton system follow desired motions and how close one could get to desired motions in the case of given actuator constraints. In addition, free design parameters of the exoskeleton can be adjusted in order to reduce the effort for desired motions.