Human-Robotic Variable-Stiffness Grasps of Small-Fruit Containers Are Successful Even Under Severely Impaired Sensory Feedback

Development of Multi Joint Gripper That Achieves Transition From Pinching to Envelope Grasping With Simple Control

Volume 5B: 41st Mechanisms and Robotics Conference ◽

10.1115/detc2017-67666 ◽

2017 ◽

Author(s):

Takumi Tamamoto ◽

Soichiro Nomura ◽

Keita Takeuchi ◽

Koichi Koganezawa

Keyword(s):

Sensory Feedback ◽

Variable Stiffness ◽

Gear System ◽

The Other ◽

Differential Gear

This study proposes a Multi Joint Gripper (MJG) that achieves envelope grasping for unknown shape objects with simple control. In our previous study we have developed the MJG having a number of multi joint fingers. Each finger has back drivability with Variable Stiffness Mechanism (VSM) for adjusting the stiffness of joints. The MJG succeeded envelope grasping of various shape objects with no sensory feedback. We have also showed the MJG is able to handle objects by pinching with the finger tips and to shift into an envelope grasping. However, in some tasks requiring transition from pinching to envelope grasping, it occasionally failed because of difference of mode setting between pinching and grasping. In this paper, we discuss the pinching motion and the transition from pinching to envelope grasping. The finger mechanism we have proposed is mainly composed of a serially connected Differential Gear System (DGS), and it is controlled by only two actuators. One for driving all of the joints simultaneously and the other for VSM of every joints all together. We propose a novel mechanism for transition from pinching to envelope grasping , which requires no change of mechanical mode setting of the DGS. The experiments revealed that the MJG successfully achieves transition from pinching to envelope grasping with very simple control.

Download Full-text

Human Cerebral Potentials During Motor Training Under Different Forms of Sensory Feedback

Journal of Psychophysiology ◽

10.1027//0269-8803.13.4.234 ◽

1999 ◽

Vol 13 (4) ◽

pp. 234-244

Author(s):

Uwe Niederberger ◽

Wolf-Dieter Gerber

Keyword(s):

Healthy Subjects ◽

Sensory Feedback ◽

Motor Task ◽

Tactile Feedback ◽

Proprioceptive Feedback ◽

Motor Training ◽

Feedback Group ◽

Emg Feedback ◽

The Right ◽

Training Success

Abstract In two experiments with four and two groups of healthy subjects, a novel motor task, the voluntary abduction of the right big toe, was trained. This task cannot usually be performed without training and is therefore ideal for the study of elementary motor learning. A systematic variation of proprioceptive, tactile, visual, and EMG feedback was used. In addition to peripheral measurements such as the voluntary range of motion and EMG output during training, a three-channel EEG was recorded over Cz, C3, and C4. The movement-related brain potential during distinct periods of the training was analyzed as a central nervous parameter of the ongoing learning process. In experiment I, we randomized four groups of 12 subjects each (group P: proprioceptive feedback; group PT: proprioceptive and tactile feedback; group PTV: proprioceptive, tactile, and visual feedback; group PTEMG: proprioceptive, tactile, and EMG feedback). Best training results were reported from the PTEMG and PTV groups. The movement-preceding cortical activity, in the form of the amplitude of the readiness potential at the time of EMG onset, was greatest in these two groups. Results of experiment II revealed a similar effect, with a greater training success and a higher electrocortical activation under additional EMG feedback compared to proprioceptive feedback alone. Sensory EMG feedback as evaluated by peripheral and central nervous measurements appears to be useful in motor training and neuromuscular re-education.

Download Full-text