Leveraging anteroposterior force oscillations to assist walking.

Although humans employ various strategies to minimize metabolic cost, walking still has a substantial cost. Half of this cost comes from leg work to accelerate and decelerate the center of mass (CoM) in the anteroposterior direction. Studies showed how applying forward forces to assist CoM acceleration drastically reduces metabolic cost. However, the effect of combining this with assisting deceleration remains unclear. Here we show that oscillating forward and backward forces with zero net work in total can also reduce the metabolic cost by 11%. In the first part of this manuscript, we used a robotic tether to apply timed forces and found forward forces are most effective when applied around the middle of the double stance (approximately 0-30% of the step cycle), whereas the optimal window for backward forces was broader, spanning roughly 50-80% of the step cycle, and the effect of timing was not strong for this force direction. Nevertheless, the results suggest that a passive device, such as a mass-spring system, could also reduce metabolic cost. In the second part, we describe simulations and experiments of walking with a wheeled mass-spring system. Simulations suggest that certain stiffnesses could reduce the estimated metabolic cost. However, human experiments showed around 60% increase in metabolic cost rather than a reduction, likely due to friction and suboptimal timings. The finding that the results did not match theoretical predictions is similar to challenges in other wearable robotics studies.