Adding low-cost passive toe joints to the feet structure of SURENA III humanoid robot

Abstract

SUMMARYAdding active toe joints to a humanoid robot structure has lots of difficulties such as mounting a small motor and an encoder on the robot feet. Conversely, adding passive toe joints is simple, since it only consists of a spring and a damper. Due to lots of benefits of implementing passive toe joints, mentioned in the literature, the goal of this study is to add passive toe joints to the SURENA III humanoid robot which was designed and fabricated at the Center of Advanced Systems and Technologies (CAST), University of Tehran. To this end, a simple passive toe joint is designed and fabricated, at first. Then, stiffness and damping coefficients are calculated using a vision-based measurement. Afterwards, a gait planning routine for humanoid robots equipped with passive toe joints is implemented. The tip-over stability of the gait is studied, considering the vibration of the passive toe joints in swing phases. The multi-body dynamics of the robot equipped with passive toe joints are presented using the Lagrange approach. Furthermore, system identification routine is adopted to model the dynamic behaviors of the power transmission system. By adding the calculated actuating torques for these two models, the whole dynamic model of the robot is computed. Finally, the performance of the proposed approach is evaluated by several simulations and experimental results. Results show that using passive toe joints reduces energy consumption of ankle and knee joints by 15.3% and 9.0%, respectively. Moreover, with relatively large values of stiffness coefficients, the required torque and power of the knee and hip joints during heel-off motion reduces as the ankle joint torque and power increases.

Online adaptation for humanoids walking on uncertain surfaces

“…Furthermore, direct joint position control yields good disturbance rejection and compensation of gear friction with high bandwidth. In order to examine the effectiveness of the proposed approach, a pattern for walking on a flat surface is generated 29 . Then, during implementation of the generated walking pattern, two flat obstacles with different heights are located in the robot path.…”

“…As it can be observed, the pelvis in the sagittal direction moves one stride in each walking cycle. Moreover, boundary conditions of the pelvis for transition between the SSP and the DSP are considered such that provide an optimal pattern 29 . In the lateral direction, a symmetric motion is considered to preserve feasibility of motion in this direction.…”

“…The topology that is adopted in this article for walking considers the feet parallel to the ground surface during walking, without heel-off or toe-off phases. 29,30 Therefore, orientation of the feet remains constant, parallel to the ground surface, in all phases of the motion. In this topology, during the double support phase (DSP), both feet are completely in contact with the ground surface.…”

“…Since many parameters may be considered that satisfy these conditions among all patterns, an optimization routine can be exploited for selecting these parameters consistent with the constraints. 29…”

Effect of Toe Length on Biped Walking Behavior

“…In particular, Kouchaki and Sadigh have considered the effect of toe-joint bending on biped gait performance [4]. Sadedel et al added low-cost passive toe joints to the feet structure of the SURENA III humanoid robot; using passive toe joints reduced the energy consumption of ankle and knee joints in comparison with a similar toe-less robot [5]. In addition, a humanlike ankle-foot complex proposed by Narioka et al [6] is designed to imitate the truss mechanism and the windlass mechanism of human beings.…”