Articulated legs allow energy optimization across different speeds for legged robots with elastically suspended loads

Abstract

Legged robots are a promising technology whose use is limited by their high energy consumption. Biological and biomechanical studies have shown that the vibration generated by elastically suspended masses provides an energy advantage over rigidly carrying the same load. The robotic validation of these findings has only scarcely been explored in the dynamic walking case. In this context, a relationship has emerged between the design parameters and the actuation that generates the optimal gait. Although very relevant, these studies lack a generalizable analysis of different locomotion modes and a possible strategy to obtain optimal locomotion at different speeds. To this end, we propose the use of articulated legs in an extended Spring-Loaded Inverted Pendulum (SLIP) model with an elastically suspended mass. Thanks to this model, we show how stiffness and damping can be modulated through articulated legs by selecting the knee angle at touch-down. Therefore, by choosing different body postures, it is possible to vary the control parameters and reach different energetically optimal speeds. At the same time, this modeling allows the study of the stability of the defined system. The results show how suitable control choices reduce energy expenditure by 16% at the limit cycle at a chosen speed. The demonstrated strategy could be used in the design and control of legged robots where energy consumption would be dynamically optimal and usage time would be significantly increased.