Intelligent System Control

Research Staff

  • Prof. Kenji Sugimoto

    Prof.
    Kenji Sugimoto

  • Assoc.Prof. Takamitsu Matsubara

    Assoc.Prof.
    Takamitsu Matsubara

  • Assist.Prof. Taisuke Kobayashi

    Assist.Prof.
    Taisuke Kobayashi

  • Assist.Prof. Masaki Ogura

    Assist.Prof.
    Masaki Ogura

  • Assist.Prof. Yunduan Cui

    Assist.Prof.
    Yunduan Cui

E-mail { kenji, takam-m, kobayashi, oguram, cui.yunduan.ck4 }[at] is.naist.jp

Research Area

1. Control systems design

Advanced robust/adaptive control
We study advanced theories in post-modern robust/adaptive control and their applications including current investigations into various schemes of feedforward learning control (feedback error learning). System identification and state estimation are also topics of interest.

Networked dynamical systems
The goal of this research is to provide a better understanding of dynamical processes taking place over complex networks, as well as developing effective strategies to control their behavior. Applications of this research direction can be found in a wide variety of contexts, from social networks to networked infrastructure and cyber-physical systems.

Distributed control
We conduct theoretical and experimental studies on distributed LED lighting systems. We also study power demand-supply balance control for distributed generation network systems composed of multiple generators such as gas-engine and photovoltaic generators.

2. Machine learning for robotics

Motor skill learning for humanoid robots
We are developing novel methods that enable robots to learn complex motor skills (e.g., biped walking, putting on T-shirts and clothing assistance) by optimal control and reinforcement learning.

Truly autonomous robot
The ultimate goal of this research is to develop the next generation autonomous robot that autonomously finds multiple objectives, selects what the robot wishes to achieve from among them, and acquires dynamic motions to achieve the selected objective.

Constructing practical myoelectric interfaces for robot control
We construct myoelectric interfaces robust to postural changes, sweating, and muscular fatigue, using surface electromyograms (sEMG) via modern machine learning methods.

Key Features

We welcome motivated students from various fields including mechanical/electrical engineering, mathematical/physical science, as well as computer science. The faculty guides students individually, taking into account their backgrounds, and assists them in mastering mathematical system approaches by the end of their course. Thereby they acquire a wide range of technical skills from fundamental theories to applications. The students in our lab are highly motivated, diligent, cooperative and eager to learn from others. We anxiously await such students from all over the world.

Fig.1 Feedback error learning control

Fig.1 Feedback error learning control

Fig.2 Quantized control of mechanical system

Fig.2 Quantized control of mechanical system

Fig.3 System Identification by manifold learning

Fig.3 System identification by manifold learning

Fig.4 Motor skill learning by enforcement learning

Fig.4 Motor skill learning by enforcement learning

Fig.5 Distributed lighting system and distributed generation network
system

Fig.5 Distributed lighting system and distributed generation network system