A better understanding of how actuator design supports locomotor function may help develop novel and more functional powered assistive devices or robotic legged systems. Legged robots comprise passive parts (e.g., segments, joints and connections) which are moved in a coordinated manner by actuators. In this study, we propose a novel concept of a hybrid electric-pneumatic actuator (EPA) as an enhanced variable impedance actuator (VIA). EPA is consisted of a pneumatic artificial muscle (PAM) and an electric motor (EM). In contrast to other VIAs, the pneumatic artificial muscle (PAM) within the EPA provides not only adaptable compliance, but also an additional powerful actuator with muscle-like properties, which can be arranged in different combinations (e.g., in series or parallel) to the EM. The novel hybrid actuator shares the advantages of both integrated actuator types combining precise control of EM with compliant energy storage of PAM, which are required for efficient and adjustable locomotion. Experimental and simulation results based on the new dynamic model of PAM support the hypothesis that combination of the two actuators can improve efficiency (energy and peak power) and performance, while does not increase control complexity and weight, considerably. Finally, the experiments on EPA adapted bipedal robot (knee joint of the BioBiped3 robot) show improved efficiency of the actuator at different frequencies.

Male pufferfish (Torquigener albomaculosus) construct “mystery circles”, geometric circular structures with radially aligned peaks and valleys, on sandy seabeds for reproduction. We discovered an earliest-stage mystery circle composed of dozens of irregular depressions. The structure changed to a primitive circular form with radially aligned valleys and a circular depression in the central area on day 2. The number of valleys increased, and the radius of the structure was much smaller than that of the final structure. These features were observed from days 2 to 4, considered the early stage of construction. The mystery circle had radially aligned peaks and valleys with a central flat area, while the radius and the differences in elevation between the peaks and valleys were increasing. These features, which were observed from days 5 to 8, were considered the middle stage of construction. On day 9, the mystery circle was completed. An irregular pattern was formed in the central area, and the radially aligned peaks were decorated with shell and coral fragments. These are the first recorded observations of construction of a mystery circle from the earliest stage.

Several factors affect the performance of humanoid walking. One factor is the complex nature of lower limbs, especially the muscles around the pelvis that contribute significantly to the stability and adaptivity of humanoid locomotion. The significance of this muscle group assures a impact on the facility of walking robots once the nature of its contribution is understood, and it can be replicated on robots. To propose a mechanical structure that facilitates walking in robots, we realized a muscle by modeling its pelvis region like that of a humanoid and developing a musculoskeletal humanoid robot. Especially, we focused on the gluteus medius, which is important for the general stability against frontal movements of the hip. Furthermore, it passively changes its influence on such motions; this is helpful during the different phases of locomotion. These changes depend on the alignment of the pelvis and femur. We confirmed the viability of the robotic gluteus medius, which was simplified to a model of two partial muscles by accomplishing the walking using this robot. This accomplishment verifies our hypothesis that using this model, the supporting functionality for the locomotion of the muscle can be reproduced and enhances the biological plausibility.

This paper indicates the development of control system for musculoskeletal robots driven by pneumatic artificial muscles. This system mimics the nervous system. The nervous system, especially spinal reflexes greatly contribute to the control of fast physical interactions without the influence of higher level control systems such as brain. Therefore, to reproduce human-like motions in musculoskeletal robots, it is necessary to imitate not only the human musculoskeletal system, but also the nervous system controlling spinal reflexes. To this end, sensors that measure and encode body movements similar to human proprioceptors are needed as the starting point for creating spinal reflexes. In this study, to produce spinal-like reflexes in musculoskeletal robots using a biomimetic approach, we developed the local control system inspired by nervous system. Using this system and a robot, we produce stretch reflex motion by human-like local feedback control.

The upper body and the lower body of previous humanoid robots has been functionally separated. On the other hand, in human, there are connected by the three-dimensional structure of psoas major. And the upper body moves when the lower body moves. It assumed that the interlocking function of psoas major contribute stabilization of whole body movement while walking. In this research, we aim to confirm the interlocking function, which caused by the three-dimensional structure of psoas major which is connected the upper and lower body, by development of biped robot. At the first, we develop a musculoskeletal biped robot which has pneumatics artificial muscles which are typical muscle that drives the lower part of a human body. Next, in order to investigate the effect of the interlocking function, we make the robot an experiment of walking motion As the result, the upper body moves by interlocking with the movement of lower body because of the three-dimensional structure of a pneumatic artificial muscle which is equivalent of psoas major.