The facial expressions of humanoid robots play a crucial role in human–computer information interactions. However, there is a lack of quantitative evaluation methods for the anthropomorphism of robot facial expressions. In this study, we designed and manufactured a humanoid robot head that was capable of successfully realizing six basic facial expressions. The driving force behind the mechanism was efficiently transmitted to the silicone skin through a rigid linkage drive and snap button connection, which improves both the driving efficiency and the lifespan of the silicone skin. We used human facial expressions as a basis for simulating and acquiring the movement parameters. Subsequently, we designed a control system for the humanoid robot head in order to achieve these facial expressions. Moreover, we used a flexible vertical graphene sensor to measure strain on both the human face and the silicone skin of the humanoid robot head. We then proposed a method to evaluate the anthropomorphic degree of the robot’s facial expressions by using the difference rate of strain. The feasibility of this method was confirmed through experiments in facial expression recognition. The evaluation results indicated a high degree of anthropomorphism for the six basic facial expressions which were achieved by the humanoid robot head. Moreover, this study also investigates factors affecting the reproduction of expressions. Finally, the impulse was calculated based on the strain curves of the energy consumption of the humanoid robot head to complete different facial expressions. This offers a reference for fellow researchers when designing humanoid robot heads, based on energy consumption ratios. To conclude, this paper offers data references for optimizing the mechanisms and selecting the drive components of the humanoid robot head. This was realized by considering the anthropomorphic degree and energy consumption of each part. Additionally, a new method for evaluating robot facial expressions is proposed.
Animal locomotion in complex environments depends on the ability to adaptively regulate movement in response to substrate mechanics. Tree frogs (Polypedates dennysi), which combine jumping and adhesive capabilities, inhabit arboreal habitats with a wide range of compliant substrates. While previous studies have offered preliminary insights into their locomotion, the biomechanical mechanisms underlying their adaptability remain poorly characterized. In this study, we developed a stiffness-adjustable takeoff substrate supported by four springs, and combined it with a 3D motion capture system to analyze the jumping dynamics and kinematics of frogs across a broader range of compliant substrates. We found that energy recovery from the substrate was influenced by compliance. On the stiffest substrate, up to 50% of the stored energy was recovered during takeoff, whereas highly compliant substrates caused nonlinear damping, energy dissipation, and even takeoff failure. During takeoff, frogs generated peak normal forces up to 6 times their body weight and fore–aft forces up to 4.5 times their body weight. However, force generation showed limited adaptability to substrate mechanics, while takeoff velocity exhibited stronger adaptability to changes in compliance. These findings reveal a trade-off between substrate mechanics and jump performance. This work provides biomechanical insight into substrate preference and informs the design of bioinspired systems capable of efficient locomotion on compliant substrates.