This paper presents a novel manipulation method utilizing dynamic deformation of a flexible body with a structural anisotropy. Employing a spiral flexible body, a dynamic underactuated manipulation using its various vibrational patterns is proposed. First, the orbit of the tip of flexible body for the vibrational input to its root is theoretically derived. Subsequently, for flexible bodies with and without the structural anisotropy, structural stiffness and vibrational orbit of the tip of body are analyzed. Through this analysis, the generation mechanism of the orbit change effect according to the input frequency is revealed. Finally, the proposed method is experimentally validated. After confirming the orbit change effect in a spiral flexible body, this effect is applied to an underactuated nonprehensile manipulation where three-Dof motion of an object is controlled by a single actuator.

An image-processing-based method for quantitatively evaluating the state of a food bolus formed via human oral processing is presented. The proposed method consists of the acquisition of the bolus image and the estimation of the mastication class using a convolutional neural network (CNN). The experimental evaluation was conducted using old-fashioned doughnuts. Images of boluses masticated 0-30 times by a subject were acquired, and they were classified into four and seven classes. After preprocessing for input images, the CNN model between the input images and the mastication class was trained. The experimental results showed that the proposed method can approximately estimate the mastication class of bolus by only one local image. In addition, it was indicated that the proposed method has a higher performance in class estimation than human visual observation.

This paper proposes a dexterous nonprehensile manipulation using the vibration of a plate, in which a three-degree-of-freedom (DoF) motion of a part is controlled based on a single actuator. First, a manipulator whose end effector is a flat plate is introduced. The manipulator employs a hybrid joint mechanism with a viscoelasticity and a nonparallel axis layout. The characteristic of the mechanism is that the shape, orientation, and size of the vibrational orbit of the plate vary based on the sinusoidal displacement input to the actuator. Subsequently, the trajectories of multiple point masses on the plate are analyzed to understand the approximated three-DoF motion of a part. The simulation results reveal that the whirlpool-like characteristics of the trajectory map, which aid in the rotational and translational motions of the part, can be managed by the input frequency, offset angle, and amplitude of the sinusoidal displacement input. Based on the trajectory maps, nine primitives for manipulating the part are designed. Finally, the proposed manipulation scheme is experimentally validated using a prototype. After confirming the nine primitives in the experiment, applications to one-DoF, two-DoF, and three-DoF parts feeding tasks are demonstrated.

This paper presents a texture evaluation system for nursing-care paste foods with a biomimetic approach. To artificially reproduce human oral processing, an elastic imitation tongue is introduced to the compression test device of paste food. During the compression, the tongue is passively deformed and holds a paste sample. Such a tongue behavior varies with respect to characteristics of paste food. Based on this effect, we propose the method for extracting the cohesiveness of paste food by the pressure distribution measurement and the image processing. Then, we develop the texture evaluation system that estimates the value of human sensory evaluation from the feature value of the pressure distribution. The experimental results show that the proposed method can appropriately estimate the value of human sensory evaluation.

This paper presents an experimental study of the bio-inspired artificial skin consisting of silicone and embedded strain gages, in which silicone imitates the epidermis and dermis, and strain gages mimic corpuscles. The strain gages are embedded in silicone under different configurations like corpuscle in humans skin. Both static displacement and dynamic excitations are applied in arbitrary positions, with different magnitudes and frequencies. The responses are observed by measuring the output signals from strain gages. Comparison with FEM simulation of static displacement shows intuitive agreement. The responses to dynamic excitation in typical frequency range of human somatosensors are obtained both experimentally and with simulation of dynamic modeling. We found that each configuration has advantages and disadvantages. This paper shows how strain gages embedded in silicone will behave in response to both static and dynamic excitations, and suggests modeling and fundamental concepts to design a bio-inspired artificial skin sensor.

Viscoelastic materials are known to exhibit temporal response that changes force or displacement at the contact interface under position or force control, respectively. In this paper, we conduct experimental study using force control to explore and observe creep phenomenon in robotic grasping in order to better understand the nature of such contact interface, which has been widely used in soft robotic fingers, robotic feet, and contact surface of robotic arms. We found that the creep response under a constant external force exhibits the characteristics of exponentially increasing or decreasing temporal response. Such characteristics are similar in nature to those found in the relaxation response of viscoelastic materials when the grasping is under position control. Two different types of creep responses are found, depending on the state of grasping. Both Types I and II in creep response mirror the Types I and II in relaxation response. We also found that different loading rates under force control result in different elastic response, in addition to the temporal response. This is an interesting finding because the Fung's model postulates for an elastic response that is independent of, and can be separated from, the temporal response. The experimental results do not show such independence.