| Different types of powered exoskeletons have been developed to assist human movement. We propose a new method of exoskeleton control aimed at improving the agility of the user's leg movements. The method, called "active impedance" control, is designed to increase the user's preferred frequency of leg swing, and consequently increase the average speed of walking, without a rise in energetic expenditure. The new controller also enables higher accelerations on the limbs, for example when taking a quick corrective step.;In this work we present the theoretical foundations of a new strategy for walking assist, based on increasing the pendulum frequency of the legs. Emphasis is given to the potential reduction in metabolic cost. Then we discuss the design and control of a 1-DOF stationary exoskeleton we employed to test our active impedance control. The controller compensates the damping and weight of the exoskeleton's mechanism. It assists the user by compensating the inertia of the leg, thereby increasing its natural frequency.;The final part discusses a series of experiments we conducted with subjects using the exoskeleton. Subjects performed sequences of knee flexions and extensions, first unassisted, then aided by the exoskeleton. The trials involved executing a computer-based pursuit task. These experiments enabled us to assess two forms of lower-limb assist: inertia compensation and virtual negative damping.;Inertia compensation increased the preferred frequency of motion. As a result, subjects spontaneously increased the RMS velocity of their legs. An inverse dynamics model showed a reduction in the energetic contribution from the muscles. The assistive effect of the exoskeleton was confirmed by catch trials included in the protocol. Additionally, subjects performed moderately better in a position-tracking task, suggesting improved ability to execute movements requiring large accelerations.;Negative damping was effective at reducing muscle torques involved in leg swing, although with limited influence on kinematics. The assistive effect of negative damping was confirmed by EMG data and catch trials. Efficiency was somewhat limited because subjects tended to resist the exoskeleton's torques at certain intervals.;The final chapter discusses future directions for research in active impedance control. We propose a concept for a wearable exoskeleton designed to assist actual walking. (Full text of this dissertation may be available via the University of Florida Libraries web site. Please check http://www.uflib.ufl.edu/etd.html). |
