Stable Manipulation Mechanism of Dexterous Robot Hand based on Multi-sensor Information Fusion Technology under Uncertainty

The projects aims at transferring human manual skills to robotic systems and studies human hand and arm motions within a generalized framework that integrates multiple sensors. The genralised framework captures and integrates finger joint trajectories, hand contact forces, and the forearm electromyography (EMG) signals. The work involves sensor integration, signal preprocessing, and correlation analysis of the multimodal sensory information.

Although the true mechanisms behind human motions and other intelligent activies are still largely unknown, research in intelligent robotics has found, and made use of, some of the relationships between the articulated motions, for example, the motions around the finger joints, and the structures of senory information such as haptic, visual and EMG information. This understanding makes it possible to develop a variety of robotic applications that utilise such information with success. However, the theory and technology for developing dexterous robot systems that fuse a range of senory information and possess high level dexterity is still an open research problem. This project is a step toward that direction.