What do walking humans want from mechatronics?

Abstract

It is an exciting time to be developing robotic prostheses, exoskeletons, and gait trainers, with clever new innovations emerging at a rapid pace. But are these the droids we're looking for? It is very difficult to predict how a human will respond and adapt to forceful interactions with an electromechanical device, and many years of development are typically required before proposed designs can be tested on humans. What if we could test our ideas for device function quickly, without the overhead of designing a product-ready prototype? This might lead to faster, and more meaningful, understanding of design requirements and trade-offs for human users. We will describe a system that we have developed for rapid emulation of robotic ankle prostheses and orthoses, and present initial results from the high-throughput experiments that this technology has enabled. One set of experiments provides quantitative insights into the optimal prosthesis motor and battery size for a given user, while another set identifies the relationships between energy cost, balance, and variability during gait. Experiments with an ankle-foot orthosis demonstrate shaping of the human energy-cost landscape, revealing that least-effort drives can be harnessed to shape self-selected coordination patterns, with applications to gait rehabilitation. We think this approach will facilitate faster identification of what humans need from wearable robots, providing detailed design requirements for engineers and resulting in better assistive technologies, sooner.