Control mechanisms for postural stability and trunk motion in bipedal running : a numerical study for humans, avians, and bipedal robots

Abstract

In this dissertation, we investigate bio-inspired postural control mechanisms for bipedal running that actively use trunk oscillations to generate natural gaits. To this end, we employ a simplified template model, the spring-loaded inverted pendulum model with a trunk (TSLIP), which captures the essential characteristics of bipedal locomotion. Based on this model, we thoroughly investigate a recent bio-inspired control method called the virtual point control, which directs the ground reaction forces toward a single virtual point, and evaluate the resulting gait characteristics and energetics both in simulation and based on real-world data. We show that the virtual point control is a good candidate for achieving postural stability in running and can inspire better controllers for future legged robots and exoskeletons.