Exploiting Natural Dynamics to Reduce Energy Consumption by Controlling the Compliance of Soft Actuators

TLDR

Exploiting natural dynamics for bipedal locomotion is gaining interest because of its energy efficiency, yet fixed natural trajectories limit mobility. The study proposes a semi‑passive walker with adaptable compliance actuators and a compliance controller that continuously adjusts compliance to shape the natural motion toward desired trajectories. The controller fits actuator compliance to the natural compliance of the desired trajectory and combines this with trajectory‑tracking control, implemented on a 1‑DOF pendulum driven by antagonistic pleated pneumatic artificial muscles. Simulations and measurements show that the compliance‑selection strategy reduces control activity and energy consumption while preserving tracking precision, enabling diverse trajectories at low energy cost.

Abstract

Exploiting natural dynamics for bipedal locomotion, or passive walking, is gaining interest because of its energy efficiency. However, the natural trajectories of a passive walker are fixed during the design, thus limiting its mobility. A possible solution to this problem is creating a “semi-passive walker” equipped with actuators with adaptable compliance, which allows the natural dynamics to be changed according to the situation. This paper proposes a compliance controller, a strategy for continuously changing the compliance in such a way as to adapt the natural motion of the system to a desired trajectory. This opens up the possibility of following a range of different trajectories with a relatively low energy consumption. The idea is to fit the controllable actuator compliance to the “natural” compliance of the desired trajectory, and combine that with trajectory tracking control. This strategy was implemented and tested on a 1-DOF pendulum setup actuated by an antagonistic pair of pleated pneumatic artificial muscles. Both simulations and measurements show that the proposed strategy for choosing actuator compliance can significantly reduce the amount of control activity and energy consumption without harming tracking precision.