1.Visual Servoing

Visual servoing is a framework in which vision process and robot dynamics are combined and formulated as a unified dynamical system. Our interests include image feature extraction techniques suitable for feedback control, and robot-camera system configuration robust against camera calibration errors. We also provide a visual tracking method that is robust to deformation and a visual servoing method that is robust to occlusion.

2. 3D Data Processing

Development and application research on 3D vision systems is being advanced with the aim of processing 3D data measured by 3D sensors in real time and applying it to robot control. 
In addition, based on research on basic technologies for 3D data processing using deep learning and topology, we are applying 3D data processing technologies to fields other than robotics (crystal structures in materials science, shape generation in computational engineering, etc.).

3.System Ethological Application

Ethology is the study of animal behavior. Our objective is to analyze the neural system that cause behavior change using model animals like nematoda (C. elegans) or fruit flies (Drosophilidae). We are developing systems for observing and analyzing the neural activities in these animals using high-speed vision and robotics.