Kazuo Hongo

An advanced musculoskeletal humanoid Kojiro

We have been promoting a project of musculoskeletal humanoids. The project aims at the long-term goal of human-symbiotic robots as well as the mid-term goal of necessary design and control concepts for musculoskeletal robots. This paper presents the concepts and aim of the project and also shows the outline of our latest results about development of new musculoskeletal humanoid Kojiro, which is the succeeding version of our previous Kotaro.

Development of bilateral wearable device “kento” for control robots using muscle actuator modules

A method of giving intelligence to control robots by human navigations is one of appropriate solution to respond to a lot of unforeseeable situations, and making movement of complicated robots in contact condition with surroundings is not easy. In this research, a bilateral wearable device using stiffness adjustable muscle actuator modules is developed in order to control robots, feel external force of robots and teach motion that contact on surroundings. For an experiment it was realized maneuvering a musculo-skeletal humanoid, it could feedback humanoids external force to manipulator and its force feedback by bilateral control leads manipulator to do safety contact motion.

Realization of large joint movement while standing by a musculoskeletal humanoid using its spine and legs coordinately

We are developing the novel musculoskeletal humanoid that is able to do coordinated motion with arms, legs and spine. In this research, we present how to realize largely bending/extending motion while standing using spine and hip joints of a fullbody musculoskeletal humanoid, which has a complicated body like a human, such as over 90 muscles and about 60 DOFs. And we propose how to generate coordinated motions of spine and hip joints and how to control such a complicated humanoid in order to keep standing during its motion. Furthermore, we confirmed that this robot could bend/extend its spine and hip joints coordinately while standing by some experiments.

Picking up dishes based on active groping with multisensory robot hand

In this paper we present a new method for picking up dishes based on active groping. Though a birds eye view is commonly used to recognize dishes, this method tends to produce errors in the presence of large occlusion, A multisensory robot hand can be used in a kitchen environment to probe and grasp dishes that are placed close to each other. Though sensing is an effective method, it is difficult to recognize a dish globally by this method because of the locality of measurement. To solve this difficulty, we propose a sensing method to aquire the geometric information about dishes by tracing their surface. We demonstrate that a robot can pick up three types of dishes on the basis of this active groping strategy by conducting an experiment.

Design of the musculoskeletal trunk and realization of powerful motions using spines

We are developing the novel musculoskeletal humanoid that is able to do coordinated motion with arms, legs and especially trunk body. In this research, we designed and implemented the trunk consisting of redundant articulated spine joints which can hold up and drive the upper body and pelvis. Furthermore, we confirmed that this trunk body ability to do full body motion, by an experiment of spine rotation motions with the real humanoid developed.

Automatic parameter adjustment of reflexive walking of a musculo-skeletal humanoid

Humanoid robots need mechanical flexibility, redundancy, multiple degrees of freedom, and so on. Because the hard and heavy body can hurt human or surrounding objects, multiple degrees of freedom and flexibility are needed in a daily life. A novel musculo-skeletal humanoid dasiaKojiropsila is one of the answer to these demand. It has those physical characteristics. How to move this complex bodily structure is a challenging theme. Especially motion generation using only partial information by itself make actional multiplicity, it will be useful in diversified environment. This paper presents the approach to make Kojiro walk and make Kojiro adjust self-parameter to walk using 6-axis force sensors in the feet.

Muscle geometric topology estimation based on muscle length - joint angle nonlinearity in tendon-driven robot systems

Many tendon-driven systems have been developed to realize more natural and various motions, or to lighten the end of a robot arm. However, the control based on the geometrical model, one of control methods the main of recent, is not easy for tendon-driven robots because more information is necessary, for instance muscle layout etc. This paper describes a muscle layout estimation methods for tendon-driven robots. This method based on nonlinearity between relative displacement of muscle length and relative displacement of joint angle. Furthermore, the confirmatory experiment was done using the simulator and the real musculo-skeletal humanoid.