In order for robots to be able to assist humans at a very close distance, robots should allow contacts occurred at many places and deal with them. For realizing such functions, robots should have whole-body soft sensor exterior for preparing contacts against almost every body parts since it is difficult to limit where to be touched from humans. Although several groups have developed humanoid type robots which have soft tactile sensors, most of them detects distributed 1-axis forces and that is not sufficient when the robot and a human have contacts very closely. In addition to that, there is no consideration in the previous studies about humans' movement during close contact with the robots. In order to solve these problems, we have newly developed a humanoid with soft 3D-deformable sensor flesh and automatic recoverable mechanical overload protection mechanism. Soft 3D-deformable sensors are implemented by molding the infrared light receiving devices into the urethane cube and by detecting the changes of the output voltage of the devices during the deformation. On the other hand, automatic recoverable mechanical overload protection mechanism is implemented by small mechanical torque limiters and monitoring system for embedded potentiometers in the torque limiters. Also detail implementation of the embedded electric system and the overall software structure is described in this paper.

In some important robotic human daily life support applications, such as housework or nursing care, soft sensor exterior is needed for allowing close contacts against humans or environments. Although there are many studies on tactile sensing elements themselves and some of them realize the mechanical softness of the element, most of their studies just propose their detection theory with their sensor element prototypes and it is difficult to use them for developing soft tactile exterior. Therefore a soft sensor exterior embedded with multi-axis deformation sensor is developed in this paper. The most important feature for this sensor is that it can detect multi-axis deformation, and the sensor element itself has enough softness to deform with the outer exterior. The developed sensor is constructed by molding a soft urethane foam with infrared LEDs and corresponding phototransistors. It detects 3D deformation of the sensor as voltage changes at each phototransistors. Furthermore, a prototype of soft sensor exterior embedded with the developed sensing elements is realized and we confirmed that the 3D deformation, such as stroking, pinching, and pushing, can be detected using the developed sensor exterior prototype.

In this paper, the development of a robot which has "flesh" made of soft polyurethane foam is presented. In order for a robot to behave in contact with an environment, people and itself, it is necessary to have not only rigid bone but also soft flesh. We developed a small humanoid robot "macra" which has thick soft flesh with distributed tactile sensors. A comparative discussion about possible materials of soft cover is described, and then thermal design was explored. A problem about movability of joints and implementation of tactile sensing was described. In the last part of this paper, several examples of advantages which are derived from soft polyurethane foam exterior parts with distributed three-axis force/torque sensors are shown.

We are trying to develop an innovative humanoid which learns from the structure of a human shoulder, especially the bladebone and collarbone. We consider that the bladebone-collarbone structures have the following merits, l)wide range of movement, 2)having the shoulder structure built on to the outside of the chest, allowing space for a lot of parts inside the chest, 3) flexible shoulder like humans. In contrast, it is difficult to control the shoulder structure because it have multiple degrees of freedom and extraordinary physical constraints. An aim of our research is to create a repeatable motion generation system of the bladebone-collarbone structures for real musculo-skeletal humanoid robots. In this paper, we describe the design process of the bladebone-collarbone structures for a muscle-driven humanoid robot, and the simulation system considering geometrical constraints of the structures. Then, we show a motion experiment using the shoulder structures of the humanoid, based on the simulation system we presented.

Shape-memory alloy(SMA) is a suitable actuator for a micro mobile robot due to its size and weight. However, length of SMA actuators is known to be nonlinear and hysteretic. We can control SMA’s shape analogously to its electrical resistance. This paper presents a platform to construct high degree of freedom robots with tendons of SMA actuators. We incorporate coil shape SMA as variable length tendons of tendon-driven robots. By controlling electrical resistance of SMA tendons, we can control posture of the robot structure in model-based method. A method to resolve electrical resistance for control from given posture of a high degree of freedom robot by using model of SMA and the robot is proposed. A small electrical system for high D.O.F. robots is also developed. As an application of our system, we show a four-legged robot with 16 D.O.F. joints and 24 SMA tendons.

This paper presents the concepts, issues on development, and future perspectives of a novel musculoskeletal humanoid `Kotaro'. The specifications and advantages of Kotaro include muscle-driven endoskeletal structure, multiple degrees of freedom, variable physical softness, the multiple-joint spine, easily configurable muscles, distributed onbody controllers, various sorts of many sensors including eyes, ears, and whole-body tactile sensors, and so on. These characteristics would achieve a flexible body, litheness of motions, safety, and adaptability and applicability to diverse tasks which would very often appear in human's daily lives

We have been studying about muscle-tendon spined humanoid robots for aiming a novel humanoid robot that has modifiable mechanical flexibility, safety and wide variety of postures. This paper describes a new concept of reinforceable-muscle humanoid and the development of a new musculo-skeletal humanoid robot named "Kenji" that is a prototype muscle-driven humanoid next to previous developed humanoid "Kenta". Kenji has 140 actuators in total and numerous multi-modal sensors. This paper also presents some experiments that show the feasibility of the robot.

We are trying to create an innovative humanoid which has increased flexibility by imitating the structure of a human shoulder, especially the shoulder blade. We designed a flexible shoulder structure for a muscle-driven humanoid robot which has the following merits, 1) wide range of movement and flexible shoulder like humans, 2) having the shoulder structure built on to the outside of the chest, allowing space for a lot of parts inside the chest. We developed a prototype humanoid robot named "Blade" which has a musculo-skeletal humanoid torso with the shoulder blade, driven by motor-actuated tendons. This paper describes the advantages of the structure of the shoulder blade, the development of a humanoid with the shoulder blade, and some experiments showing its range of movement using the prototype humanoid robot we presented.

We propose a concept of reinforceable-muscle humanoid with hyper parallel muscle-tendon systems. By the facilitation of changing the assignment of actuators, we can easily strengthen a specific part of robot body, while changing the arrangement of actuators of current humanoid robots is quite difficult. By developing muscle units in each of which a motor and sensors are integrated, and by adopting musculo-skeletal structure, the rearrangeability of muscles has been realized. This paper describes the concept, prototype design and implementation of the reinforceable muscle humanoid, and the evaluating experiment by a musculo-skeletal humanoid is presented.

Since an environment or body states of robots are very changeable, robotic imitation systems should have ability to develop their behaviors by themselves. For the development of the imitation behaviors, we assume that attention during imitation is the key information. In order to realize such imitation systems with evolving ability, the idea of reinforcement learning system based on the attention structure during imitation has been presented in this paper. First, for describing the attention during imitation behaviors, we define the term 'sensor-action attention pair' as the pair of the most important sensor information and the focused body parts during that behavior. Second, we introduce R-learning, the reinforcement learning method for continual tasks such as imitation behaviors. Third, the method to design the state-action space and the reward function based on the sensor-action attention pair is proposed. At last, for the confirmation of the function of the proposed imitation behavior system, we have done some experiments using actual humanoid Kenta. In those experiments, Kenta can develop imitation behavior that imitates the hand position of the human.

It is very important and hard to develop motions and behaviors for robots that have complex body and numerous actuators and sensors. The more complex a humanoid's body becomes, the more important the developing environment of the motions and behaviors will be. We have built a fully muscle-driven flexible-spine humanoid named "Kenta", which has 96 actuators and more than 400 sensors including tension sensors, posture sensors, vision sensors, and so on. This paper describes the developing environment of motions and behaviors for the robots like Kenta. A visualizable geometric model, a simulation environment transparent to the real environment, sample procedures of the motion development, and, as an example, realization of swing behavior by Kenta are presented.

We are trying to realize a humanoid which has flexibility. If humanoids have a flexible structure, safety and posture variety can be achieved. We focus on the role of the human spine and muscle-driven system. By having a flexible spine, a humanoid will have safety and many degrees of freedom to realize a variety of postures. By driving joints by tension-controllable tendons, flexibility of the joints can be controlled. We developed a whole-body tendon-driven flexible-spine humanoid named "Kenta". This paper describes the design and control of Kenta, focusing on the design of the spine. The spine consists of ten joints, vertebrae and rubber disks, ribs, and forty muscles equipped with tension sensors. We also propose control methods of the spine. One uses a geometric virtual robot model and another is based on direct teaching. Using these methods, some whole-body motions are presented.