Variable Compliant Feet for Humanoid
The legged robots were brittle and fragile when performing locomotion tasks, such as jumping or running, where impact forces can increase significantly and reach levels of an order of magnitude. These can effectively damage the leg/foot, its actuation, and structure. I studied the effect of compliance on a leg/foot system that incorporates intrinsic elasticity both at the joint level and the foot, and how configuration-dependent Cartesian inertia and joint-level stiffness and damping affect the impact. Based on this, a landing motion strategy is proposed to reduce the impact force. To determine the landing motion property of multi D.O.F leg, I used a nonlinear contact model and multi-body dynamics simulation and considered a floating base model. The results demonstrated that a simple passive mechanism with a non-parallel foot posture during landing can significantly reduce the impact forces while the foot is landing. The storing energy is also maximized with edge (forefoot) landing, which means that the forefoot contact landing motion can absorb external force more efficiently. The relationship between impact force and foot stiffness/leg configuration was demonstrated. In the case of a stiff foot, the ground reaction force only depends on configuration-dependent inertia and damping coefficient at operational space. In the case of a soft foot, the configuration-dependent Cartesian stiffness of the leg becomes lower, and the difference in stiffness amplitude becomes less significant. This initial study suggested a better landing posture with the non-parallel foot posture, although it is difficult to maintain balance during foot contact. Read the reference 3.
A different approach was involved to overcome the balancing problem from the previous study, which is keeping the foot horizontal to the flat ground. From the conventional approach to the design, in contrast to the simple planar rubber pad foot sole that is conventionally used in humanoid robots, I introduced a new foot sole design in which the dissipation of energy during collision is done effectively using a viscous air damping sole mechanism that provides better reduction of the ground impact forces. I presented the principle of the foot sole and provided details of its design and implementation. Experimental trials were performed with the child-sized humanoid robot, COMAN, wearing the proposed feet to validate their performance during landing and walking. The results demonstrate that the proposed new passive damping mechanism can effectively reduce the ground reaction impact forces and oscillations while maintaining the foot/body posture. However, the viscous air damping sole mechanism did not sufficiently provide with enough stiffness variation for walking locomotion. The foot sole requires a change in the stiffness at the beginning/during/stopping of the walking. Read the reference 2.
Targets
- Design humanoid feet that can tolerate an impact transmitted,
- Design humanoid feet that can improve the power consumption.
Role
Project lead, Research and mechanical engineering
Achievements
From the previous two research and design studies, the general approach to humanoid feet design has to consider the use of soft sole structures that are variably stiff and compatible with flat terrain locomotion. To improve humanoid feet, I presented a mechanism for new variable compliant humanoid feet which can provide functionality for humanoid locomotion in Figure 1. The proposed feet design introduces a new toe mechanism in the feet with variable stiffness implemented using a leaf spring and rubber balls in series. I presented the mechanism design and discussed the estimation of the variable stiffness range and coefficient of damping at the sole. A prototype of the feet was built as shown in Figure 2, and experimental results were presented to validate the feet design. It was designed for a child-sized humanoid robot. A stiffness estimation method was presented when a linear and a non-linear spring are connected in series. The material properties of the rubber ball sole are determined by experiments and numerical simulation. Read the reference 1.
|
||
Figure 1. CAD model of the variable compliant feet
|
Additional experimental results were conducted with iCub legs to validate the improvement of the foot mechanism's function, as well as power consumption during walking. The soft and stiff flat soles were compared using different materials. Comparing the trials, the power consumption was compared as the soft sole allowed for 18% less consumption than the stiff sole as shown in Figure 3. The walking performance was unstable and couldn't maintain a straight gait with the soft sole, while the stiff sole allowed for stable walking. The variable compliant feet with soft/stiff toe configuration exhibited similar behavior to the soft/stiff sole material experiments. Among the different stiff configurations, the difference in power consumption was 8.4%. This is presumably due to the weight of the foot mechanism. By combining the advantages of the stiff sole, which provides stable walking balance, and the soft sole, which offers lower power consumption, a stable and energy-efficient walking pattern is achieved as shown in Figure 4 and the video clip.
|
|
Figure 3. Comparison between the soft/stiff soles |
|
|
Figure 4. Experiments with the soft/stiff toe configuration |
Refereces
- Design of a Variable Compliant Humanoid Foot with a New Toe Mechanism. Robotics and Automation (ICRA), 2016 IEEE International Conference on, pp. 642 – 647.
- A New Foot Sole Design for Humanoids Robots based on Viscous Air Damping Mechanism. Intelligent Robots and Systems (IROS), 2015 IEEE/RSJ International Conference on, pp. 4498 – 4503.
- How leg/foot compliance and posture affects impact forces during landing. Robotics and Biomimetics (ROBIO), 2014 IEEE International Conference on, pp. 961 – 967.
