Nagoya University · Engineering
Professor Yoshitaka Adachi's research lab specializes in advanced materials science and intelligent mechatronics, focusing on the microstructure-property relationships in functional alloys—particularly Ti-Ni shape memory alloys and low-carbon steels—through experimental and computational approaches. The lab also pioneers haptic interface technologies for immersive human-machine interaction, integrating control systems with real-time force feedback. Recent work emphasizes data-driven materials design using machine learning and deep learning for microstructure classification and inverse materials optimization.
Figures are computed from collected data and may differ slightly.
A method of intermediate space for controlling haptic interfaces is characterized by updating a virtual plane at a low frequency while maintaining a high update rate at force control loop of the interface. By using the virtual plane, the detection of collisions between the tip of finger and virtual objects became independent from the control of the haptic interface. This will enable the haptic interface to display more and more complex surfaces in keeping the same sampling frequency of impedance
A study has been made of the effect of hydrogen on the shape memory effect and transformation behavior of Ti-Ni alloy. The material was evaluated in the temperature range from 300 to 473 K using a bent specimen. The transformation behavior was analyzed by differential scanning calorimetry (DSC) and X-ray diffraction technique.The results obtained are as follows:(1) The shape recovery rate is markedly decreased by a small amount of hydrogen which does not cause the increase in hardness.(2) The sh
Abstract In response to modern materials research, a data‐driven properties‐to‐microstructure‐to‐processing inverse analysis is proposed for use in material design. In the present work, machine learning optimization algorithms of Bayesian optimization, genetic algorithm, and particle swarm optimization are used to perform inverse analysis with a maximum property search. The use of machine learning algorithms readily involves careful tuning of learning parameters, which is often carried out by a
Deep learning by convolution neural network (CNN) was applied to recognize a microstructure of steels. Three typical CNN-models such as LeNet5, AlexNet, and GoogLeNet were examined their accuracy of recognition. In addition to a model, an effect of learning rate, dropout ratio, and mean image subtraction on recognition accuracy were also investigated. Through this study, the potency of deep learning for microstructural classification is demonstrated.
It is well established that the ferrite grain size of low-carbon steel can be refined by hot rolling of the austenite at temperature below the nonrecrystallization temperature (Tnr). The strain retained in the austenite increases the number of ferrite nuclei. In present study a C-Mn steel is heavily deformed by compression at temperature below the determined Tnr for this steel during accelerated cooling. Compression experiments are carried out at various cooling rates before deformation and temp
In this paper, a haptic device for multi-fingers is proposed. The feature of this device is as follows. (1) This device consists of a probe, several sets of force controlled manipulator, and a trajectory controlled manipulator. (2) The bases of the force controlled manipulators with small links of low inertia are attached to a tip of the trajectory controlled manipulator. The tips of force controlled manipulators are attached to an operator's fingertip, respectively. (3) The base of the probe is
Abstract To understand the material paradigm, data‐driven material design necessitates both microstructural input and output in the form of visual images. Therefore, generative adversarial networks (GAN)‐based deep convolutional GAN, cycle‐consistent GAN, and super‐resolution GAN techniques are used to generate, translate, and improve the quality of microstructural images in this study. The reconstructed virtual microstructural images are realistic and indistinguishable from the real ones. Furth
Open papers in the app to read, cite, and organize with AI.
