Establishing thresholds for swing transparency at the knee during gait to inform exoskeleton design

Abstract

Knee exoskeletons have been developed to assist, stabilize, or improve human movement or recovery. However, exoskeleton designers must implement transparency (i.e., get out of the way) modes during the swing phase of locomotor tasks to avoid impeding movement. The problem is that it is not understood how sensitive people are to small knee torques or what level of knee impedance is acceptable (sufficiently transparent) during swing phase. Here, we (i) characterized the biomechanical consequences of knee stiffness and damping during swing, and (ii) leveraged user perceptions of being impeded and toe clearance to define transparency thresholds, below which the participants were sufficiently unimpeded during the swing phase of gait. We conducted a series of human subject experiments that involved walking and stair ascent/descent while wearing a modified knee brace with five stiffness values ranging from 0 to 4 Nm/rad and five damping values ranging from 0 to 0.77 Nm/rad/s. We measured changes to lower limb kinematics, knee flexor muscle activity, and participants’ perception of being impeded during swing. Kinematics, muscle activity, and perceived impedance all changed in response to added stiffness and damping. For stiffness, we found the median transparency thresholds for walking and stairs to be 1.76 Nm/rad and 2.95 Nm/rad, respectively, which corresponds to peak knee moments during swing of around 2.3 and 5 Nm. For damping, we found the median transparency threshold for walking and stairs to be about the same, 0.29 Nm/rad/s, which corresponds to peak knee moments during swing of around 2.3 Nm. These values provide useful benchmarks for defining quantitative design requirements for knee exoskeletons intended for locomotor activities.