English UT







Joining Lab





Year: 2011-
Takahiro Ohba
Hideki Kadone
Kenji Suzuki

- Cybernics
- Augmented Human

VSLink - Variable Stiffness Link
Magneto-rheological fluid based Link System for the upper limb assistance


Robots with variable stiffness characteristics are necessary in scenarios where robots cooperate with humans while colliding with outside objects. The mechanism should be rigid when they need to transmit force and flexible when they should ease force.

We propose a novel method to change the stiffness of a mechanical link system which is assumed to be rigid originally, instead of changing the stiffness of the actuator part. This paper proposes a novel variable stiffness link system by combining the developed link and an existing motor link system. In this study, we aim to apply the developed link to upper limb assistance. Also, we are going to verify the effectiveness of the developed link by conducting several experiments to evaluate its characteristics and by applying it to upper limb orthosis.

Generally, the term stiffness refers to the difficulty of displacement against load, and equivalently. In case of high stiffness the load is difficult to displace and in the case of low stiffness the load is easy to displace. This study shows the development of a link with magneto-rheological (MR) fluid and a compression coil spring, which can change its stiffness according to the given magnetic field controlled by the current in the electrical coil.

MR fluid is a kind of smart fluid. When the fluid is not in a magnetic field, it shows the characteristics of Newton fluid. However, when the fluid is in a magnetic field, it increases its viscosity and shows the characteristics of Bingham fluid, where the flow velocity is zero when the given stress is under a certain yield point. Note that the coil current changes the yield strength. This study shows the development of a variable stiffness link (VSLink) by changing this yield strength which is modulated by the magnetic field controlled by the current on the electrical coil.


This work is partly supported by FIRST Program: World leading Human-Assistive Technology.

Related Projects



  © 2005-2011 Artificial Intelligent Laboratory, University of Tsukuba, Japan