Information Physics and Computing Faculty List

Aiming to understand, process, and control phenomena related to sound media, we will conduct research on the creation of new signal processing that is aware of wave fields and on the construction of information processing systems that apply such signal processing. Specifically, we aim to make engineering contributions to the expansion of human sound information processing capabilities and the creation of new art forms through the following research on statistical mathematics and machine learning-based signal processing.
1. communication enhancement based on acoustic signal processing:
Using statistical approaches, we will realize a flexible blind signal processing system that does not require pre-supervised information. We will also apply this approach to human interfaces and universal communication support systems.
2. music signal processing and sound augmented reality:
We will apply machine learning methods to a variety of sound media to realize high-quality music information processing systems, such as signal analysis based on spatio-temporal frequent patterns. In addition, we will construct a sound augmented reality system by combining this processing with stereoscopic sound reproduction based on wavefront synthesis theory.
3. Mathematical analysis and sensitivity quantification of nonlinear signal processing systems:
Through higher-order statistical analysis of nonlinear signal processing systems used in speech and acoustic signal processing, we will pursue statistical estimation methods that are “aurally” meaningful to humans and construct a framework for new signal processing systems.

We investigate machine learning theories for spoken language processing for augmented human-human and human-machine communication.

By integrating measurement and control technologies for fluid drive systems and system design utilizing the characteristics of actuators, we will conduct research and development of medical systems, robotic systems, and human-machine systems useful for a healthy society with longevity, such as surgical operations and motion assist. In addition, by integrating biomedical engineering and information science, we will develop intelligent and highly functional systems and implement them in society.

1) Research on intelligent and highly functional robotic systems to assist surgical operations
2) Research on body assist systems using pneumatic rubber artificial muscles
3) Research on the application of morphological computation to state estimation and prediction problems in medical systems

We conduct research in robotics engineering, focusing on the optimal design and control of human-machine collaborative robot systems from both hardware and software perspectives. This study combines mechanical elements with diverse physical properties, such as rigid linkages, fluid drives, and flexible materials, to construct novel hardware configurations that realize high performance while harmonizing with humans. Additionally, precise control and software features such as state estimation through disturbance observers are designed and implemented by thoroughly modeling individual mechanical elements, thereby enhancing robot systems' functionality.

Direct solution of inverse problems and development of measurement structures. Applications to non-invasive measurement, non-destructive testing, human-machine interface, etc.
1) Direct algebraic solution method for inverse brain magnetic field problem: Based on head surface data, we developed a method to obtain the location of neural current sources in the brain as a solution of algebraic equations.
2) Load-integral sensor for magnetic flux leakage detection: We developed a sensor that directly measures the Fourier coefficient of the magnetic field.
3) RFID tag localization and application: We developed a method to obtain the three-dimensional position of a tag as a solution of linear equations. Applied to object management and search for avalanche rescuers.

Biorobots built from biological cells and molecules have many possible applications in regenerative medicine, drug discovery, biosensing, and so on. To systematically design such biorobots without many experiments, we develop predictive design methods for the biorobots based on mathematical and physical models of cells and molecules. In particular, by using theories of nematic liquid crystals and complex analysis, we build a design method for controlling the shape and deformation of cellular sheets and molecular robots and verify the proposed method by experiments.

Research to expand the capabilities of sensors and interface devices by incorporating new physical phenomena into information processing and transmission mechanisms, including artificial skin that applies telemetry technology to tactile elements, flexible tactile sensors that use acoustic resonance phenomena, compact biometric sensors and thermally induced strong ultrasonic integrated devices, and tactile displays that artificially create a realistic tactile sensation on human skin.

Engaged in research on human-machine interfaces with a focus on the sense of touch. Recently, he has been focusing on the emotional effects of touch, especially in the sense of touch.
People perceive the same tactile stimulus as pleasant or unpleasant depending on its context. Research on robotizing stuffed toys that users are attached to so that they can move them, research on using human skin surfaces as information input surfaces, or research on generating pleasant tactile patterns through changes in temperature and other factors.

Our research focuses on control methods for large-scale networked systems utilizing heterogeneous networks.　
The study of such networked control systems and cyber-physical systems requires the two areas of systems control and informatics to meet in new forms. We work on their fundamental characterizations to more application-oriented design methodologies.
The specific research topics are as follows:

1. Control over networks
2. Distributed cooperative control of multi-agent systems
3. Cyber-physical security of control systems

Using non-invasive brain imaging techniques such as magnetoencephalography (MEG) and functional magnetic resonance imaging (fMRI), I am investigating the neural mechanisms of sensory perception and cognition, and applying the findings to engineering. In addition, I am developing methods to non-invasively control neural information, and aiming to clarify neural activities that causally contribute to perception and behavior, with the goal of applying these control technologies in the real world. Recently, I have also been focusing on research into the role of neural oscillations, which are periodic neural activities, in information integration, and the mechanisms that give rise to individual differences in neural activity and perception/cognition.

I am conducting research to understand the brain information processing that underlies human perception and cognition by combining the methods of psychophysics, neuroscience, and mathematical modeling.

• Computational Imaging
Pioneering a new field of imaging that integrates optics and information science, we are creating new principles for optical measurement and control systems. These include innovations such as lensless cameras, single-pixel cameras, and imaging through scattering media, contributing to a wide range of fields including biomedicine, astronomy, and security.
• AI Photoinics
Advancing research and development in computing and accelerators to meet the growing demands for information communication and computation through optical physics. By leveraging the diverse dynamics of optical systems as functionalities, we aim to explore innovative architectures for decision-making and reservoir computing that utilize these dynamics.

• Neuromorphic Computing
Using the computational power of physical phenomena to process information in analog or unconventional ways, such as Reservoir Computing
• Physical Deep Neural Networks
Machine-learning inspired algorithms for training physical systems to perform computation such as physical deep learning, direct-feedback alignment and adjoint-based methods
• Physical Limits of Computation
Studying the ultimate computational limits of light-based computing systems

(1) Cyber-physical systems:
Cyber-physical systems connect everything in thephysical world to the Internet, process enormous amounts of obtained data in theinformation or cyber world, and work on the physical world. We are conductingresearch on optimization of computing to improve performance, responsiveness,power efficiency, reliability, and security by making full use of characteristics of thetarget processing task.

(2) Highly Efficient Accelerated Computing:
Significant performance improvementsare required in various types of computing, including high-performance computingand machine learning. To meet this demand, we are conducting research to improvethe speed of high-efficiency computing by coordinating and linking devices, circuittechnology, architecture, and system software across design layers. To this end, weare studying coarse-grain reconfigurable architecture and approximate computingwhich makes good use of parallelism, locality, and allowed accuracy degradation.We are also working on quantum computing based on new computing principles.

Communication technology and design optimization for cloud robotics systems: Based on ROS (Robot Operating System), we are researching a lightweight runtime environment for embedded devices, and a highly autonomous communication library for IoT systems using the functional language Elixir. We are also working on development methodologies that utilize cloud-native technologies and virtual environments for robot applications in large-scale IoT environments.

Comprehensive computing technology for IoT systems: We are researching distributed machine learning infrastructures, especially for processing allocation optimization that adapts to changes in resource and geographic information of IoT nodes, and a programming model for comprehensive representation of fairness and diversity in the AI model. In addition, we are researching a resource-permeating distributed processing platform based on a functional programming paradigm.

We are conducting research on "Information Somatics," which explores the mechanisms of the body as a physical information system based on physiological, cognitive, and physical findings. It aims to enhance the innate human sensory functions, motor functions, emotional functions, and intellectual processing abilities through measurement, communication, and control technology.

(1) Extended Body: Research on technologies that extend human input/output by integrating biometric information such as gaze, facial expressions, and heart rate, with sensory and perceptual measurement technologies such as motion prediction and intention, and intervention technologies like robot control or electrical muscle stimulation. This involves engineering research and development aimed at enhancing human capabilities and acquiring new bodily perceptions by appropriately sensing the user's intent and feeding back information about the task object to the user's body.

(2) Extended Communication: The human body and mind are inseparably related, and subjective experiences such as perception and emotions are constituted through the mediation of one's own and others' bodies. In a system that includes both self and others, this research aims to transform subjective experiences by controlling the flow of physical and cognitive information using Virtual Reality (VR), augmented reality, wearable technology, wireless technology, robotic technology, and telexistence. The goal is to socially implement support for communication among users with different attributes and preferences, aiming for the realization of super-aged societies and smart cities where diverse people thrive.

We are exploring methodologies to connect cyber-physical spaces in- and outside of the human bodies. Specifically, we work on wireless technologies that connect between the body's internal (wet) and external (dry) environments, with a focus on both hardware and software innovations. This includes a novel technology that generates ultrasound inside the body without contact by irradiating modulated terahertz waves on the body surface, and a novel beam tracking technology that compensates for wave diffraction, enabling the wireless data transmission at terabit-scale towards dynamically moving targets, whether individuals or devices. Our objective is to support human activity seamlessly, ensuring that our solutions are non-intrusive and maintain the natural physicality of the users.

Research on network protocols, wide-area distributed system architectures, and cyber security for IT infrastructures to realize highly reliable and secure communication infrastructures will be conducted. While high-speed and high-capacity communication is required, flexibility and security of communication infrastructures are also required. We will conduct research and development of communication systems that aim to achieve both of these goals.

Basic research on sleep-wake rhythms and the development of technologies that contribute to the understanding and control of the sleep-wake rhythms.
Through the different hierarchies of molecules, animals, and humans, we aim to realize the understanding and control of life phenomena using sleep-wake rhythms as a model system.

1) Establishment of whole-brain whole-cell analysis technology: Development of transparency technology, high-speed imaging technology, and image processing technology.
2) Development of a sleep measurement and analysis platform: Development of medical device-level devices using acceleration sensors and information analysis technology.
3) Modeling of sleep-wake rhythm: Modeling of sleep-wake rhythm at the molecular, cellular, tissue, individual, and population levels.
4) Creation of pharmaceuticals using chemoinformatics.