Y. Mae

Object recognition using appearance models accumulated into environment

Proposes a method of object recognition using appearance models accumulated into a RFID (radio frequency identification) tag attached to the environment. Robots recognize the object using appearance models accumulated in the tag on the object. If the robot fails in recognition, it acquires a model of the object and accumulates it to the tag. Since robots in the environment observe the object from different points of view at different time, various appearance models are accumulated as time passes. In order to accumulate many models, eigenspace analysis is applied. The eigenspace is reconstructed every time robots acquire the model. Experimental result of object recognition shows effectiveness of the proposed method.

Mobile manipulation of humanoid robots-a method of adjusting leg motion for improvement of arm's manipulability

A control method of mobile manipulation for a humanoid robot by integrated motion centered arm manipulation is proposed. This method makes the arms carry out the manipulation, and the legs assist manipulation by adjusting its motion in order to keep the arms manipulability high. Firstly, the arm tips always follow their desired position under the condition with external force by impedance control. Secondly, the shoulder is controlled so that manipulability and stability can be improved. Moreover, we define three functions; they are an average of manipulability measure, and an average and a minimum of stability margin in a locomotion cycle. We utilize their values in a locomotion cycle rather than the values of them at every moment, because manipulability and stability vary with robot motion. Then, we define the total evaluation function expressed by the sum of them. Step length and timing of step motion are determined so that the function can be optimized. The effectiveness of the proposed method is ascertained through the computer simulations.

Fundamental study on human support system using humanoid robots for creating SSR society

For creating a safe, secure and reliable (SSR) society, the concept of a human support system using humanoid robots is proposed; humanoid robots exist in our environment and protect us from danger. A basic control method for humanoid robots for generating human support motion is developed. An evaluation system of the human sense of security for coexisting robots is discussed.

Development of the basic structure for an exoskeleton cyborg system

The internal human skeleton is referred to as an endoskeleton, while the cuticles of lobsters and crabs are referred to as exoskeletons. Much research has been conducted in attempts to combine endoskeletons and exoskeletons. However, there are very few instances of research being undertaken to create a cyborg. The human endoskeleton consists of kinematic pairs that are joined together with a maximum of three degrees of kinematic freedom. These are a turning pair on a lateral-bending axis, a turning pair on a longitudinal-bending axis, and a turning pair on the central axis of a skeleton. If the strength of the muscles on a human skeleton decline, or if component parts of a human skeleton are lost, then a module-type exoskeleton designed with a maximum of three degrees of kinematic freedom would be able to contain an entire body so that it could assist the functions of the arms, hands, legs, and other body parts to complement the declining or lost physical functions (kinematic or biological). Based on this concept, it is possible to develop an exoskeleton cyborg system that behaves in much the same way that a human does in daily life.

Parts and Packets Unification for Construction Automation and Robots

This paper proposes a method of configuration of database system for parts oriented construction. In parts oriented construction, robots acquire information related to the tasks from the server through ID attached to the construction components. This system aims (1) simplification in motion planning of the construction robots using information attached to components, (2) reconfiguration of data structure for process of construction tasks. The robot performs the tasks according to local planning based on the local information. The operation server reconfigure parts information based on task result by robots so that the next robot can obtain the required information from the composed components in the next phase. We call the parts integrated information “packets”, and call our idea “parts and packets unification.” The feasibility of our proposed method is also shown in the preliminary experiments where an RFID tag attached to a component is applied for data exchange.

Development of an optimum end-effector with a nano-scale uneven surface for non-adhesion cell manipulation using a micro-manipulator

In order to realize effective micro-manipulation using a micro-manipulator system, an optimum end-effector is proposed. Cell-manipulation experiments using mouse fibroblast cells are conducted, and the usability of the proposed end-effector is confirmed. A key advantage of the micro-manipulator is high-accuracy, high-speed 3D micro- and nano-scale positioning. Micro-manipulation has often been used in research involving biological cells. However, there are two important concerns with the micro-manipulator system: gripping efficiency and the release of gripped objects. When it is not possible to grip a micro-object, such as a cell, near its center, the object may be dropped during manipulation. Since the acquisition of exact position information for a micro-object in the vertical direction is difficult using a microscope, the gripping efficiency of the end-effector should be improved. Therefore, technical skill or operational support is required. Since, on the micro-scale, surface forces such as the adsorption force are greater than body forces, such as the gravitational force, the adhesion force between the end-effector and the object is strong. Therefore, manipulation techniques without adhesion are required for placed an object at an arbitrary position. In the present study, we consider direct physical contact between the end-effector and objects. First, the design and materials of the end-effector for micro-scale manipulation were optimized, and an end-effector with an optimum shape to increase the grip force was fabricated. Second, the surface of the end-effector tip was made uneven, and the adhesion force from increasing on the micro-scale was prevented. When an end-effector with an uneven surface was used, release without adhesion was successful 85.0% of the time. On the other hand, when an end-effector without an uneven surface was used, release without adhesion was successful 6.25% of the time. Therefore, the superiority of a structure with an uneven surface was demonstrated, and high-accuracy cell manipulation without sliding during handling and without adhesion was performed successfully.

Tracking mechanical and morphological dynamics of regenerating Hydra tissue fragments using a two fingered micro-robotic hand

Regeneration of a tissue fragment of freshwater polyp Hydra is accompanied by significant morphological fluctuations, suggesting the generation of active forces. In this study, we utilized a two fingered micro-robotic hand to gain insights into the mechanics of regenerating tissues. Taking advantage of a high force sensitivity (∼1 nN) of our micro-hand, we non-invasively acquired the bulk elastic modulus of tissues by keeping the strain levels low (e < 0.15). Moreover, by keeping the strain at a constant level, we monitored the stress relaxation of the Hydra tissue and determined both viscous modulus and elastic modulus simultaneously, following a simple Maxwell model. We further investigated the correlation between the frequency of force fluctuation and that of morphological fluctuation by monitoring one “tweezed” tissue and the other “intact” tissue at the same time. The obtained results clearly indicated that the magnitude and periodicity of the changes in force and shape are directly correlated, confi...

View controls in environment presentation for robot tele-operation

Operators expect to obtain sufficient information on working environment in robot tele-operation. The information may be typically provided through vision sensors and equipment due to their large amount of data. The operator requires right view angle during manipulation and/or locomotion, since objects and even robot itself may obstruct his or her view when the view point is fixed. The paper discusses two types of view controls. One is called normal or subject-centered view which provides the same field of vision centered in the operator as obtained by our natural vision system. The other is called object-centered view where the vision field is centered at and rotated around the viewed object. The view motion is controlled by the operators head motion in every view control. The two controls are applied to tele-operated manipulation and navigation tasks by the working robot which has six limbs (usable for both arm and leg) on the simulator, and they are compared and evaluated in the aspects on how the robot efficiently walks and manipulates objects.

Adaptive relocation of environment-attached storage devices for effective knowledge-sharing among multiple robots

Proposes a method of effective knowledge sharing among multiple robots by using storage devices which can be attached to the environment. If the robots store knowledge about the environment in these devices, the robots can use the stored knowledge and reduce the cost of achieving their tasks. Since the memory size of such storage device is limited, the robots cannot store all of the required knowledge. Therefore, the robots store knowledge according to its usefulness, which is defined by the reducible cost per unit time. A memory unit that is not useful to the robot in which it is stored can be copied to a storage device and deleted from the robot. It is then available, on the device, for any robot that requires it. Then the robots can store useful knowledge over the whole environment, and the memory units are flexibly re-located according to the given tasks. A simulation result of multiple robot navigation for repetitive transportation tasks shows the effectiveness of our method.