Although physiological studies have shown evidence of phase resetting during fictive locomotion, the functional roles of phase resetting in actual locomotion remain largely unclear. In this paper, we have constructed a control system for a biped robot based on physiological findings and investigated the functional roles of phase resetting in the gait transition from quadrupedal to bipedal locomotion by numerical simulations and experiments. So far, although many studies have investigated methods to achieve stable locomotor behaviors for various gait patterns of legged robots, their transitions have not been thoroughly examined. Especially, the gait transition from quadrupedal to bipedal requires drastic changes in the robot posture and the reduction of the number of supporting limbs, and therefore, the stability greatly changes during the transition. Thus, this transition poses a challenging task. We constructed a locomotion control system using an oscillator network model based on a two-layer hierarchical network model of a central pattern generator while incorporating the phase resetting mechanism and created robot motions for the gait transition based on the physiological concept of synergies. Our results, which demonstrate that phase resetting increases the robustness in gait transition, will contribute to the understanding of the phase resetting mechanism in biological systems and lead to a guiding principle to design control systems for legged robots.

This paper addresses the control of the gait change from quadrupedal to bipedal locomotion of a biped robot. In our previous work, we developed a locomotion control system using nonlinear oscillators that generate adaptive walking behavior verified by numerical simulations and hardware experiments. We extended it to deal with the gait change from quadrupedal to bipedal by designing kinematic coordination between these gait patterns, which was only verified by numerical simulations. In this paper, we create a biped robot and improve the control system to verify the performance of the proposed control system.