Research Prosthesis2012

From Rii

Jump to: navigation, search

An Underactuated Prosthetic Hand Based on Postural Synergies

December 2011 to September 2014

Note bulb.png This work was supported by the Chinese National Program on Key Basic Research Projects (the 973 Program) #2011CB013300.

Manipulation of two rehabilitation training balls on the palm: (a) a pose of the dummy hand and (b) a pose of the prosthetic hand

It is a particularly challenging task to build and attach an anthropomorphic prosthetic hand that can replicate delicate motions of the biological original. In order to achieve this goal, the prosthetic hand shall be versatile enough for various daily tasks and controllable through a bio-signal interface, such as EMG (electro-myography) or EEG (electro-encephalography). However limited bandwidth of these interfaces used to prevent fully actuated robotic hands from being applied as prostheses if each DoF (Degree of Freedom) requires individual control to perform dexterous grasping tasks, even though many designs were absolutely the state-of-the-art.

Recent advances in neurology showed that CNS (Central Nervous System) controls hand muscles in a coordinated manner. This coordination is referred as to a postural synergy. Each postural synergy corresponds to flexion/extension actuation statuses of several involved muscles. CNS combines postural synergies, adjusting each synergy’s weight (coefficient), to realize various hand motions. Combination of two primary postural synergies accounts for more than 80% of the variance of dozens of different grasping postures.

This project proposes a complete process of designing underactuated prosthetic hands to realize intended continuous hand motions. The specific paradigm is to rotate two rehabilitation training balls on palm using coordinated finger motions as shown in the figure. This exercise helps the elderly or patients after mild stroke to maintain or recover their hand motor function. Although this motion sequence might not seem practically meaningful for amputees, the motivation here is to demonstrate the capability and effectiveness of this proposed design process.

The dummy hand manually posed for a motion sequence of manipulating rehabilitation training balls

In order to realize the intended motion sequence of manipulating rehabilitation training balls on a prosthetic hand, a dummy hand was first constructed and manually posed for six different key poses. The dummy hand had identical enveloping dimensions and geometry as the one to be constructed with transmissions and actuations. The poses of the dummy hand were measured using an optical tracker and the postural synergies were extracted to guide the design of a complicated transmission system which maps two synergy inputs to 13 outputs to drive the prosthetic hand. This transmission system is hence referred to as a mechanical implementation of the postural synergies.

Then the prosthetic hand with transmission was designed, following the exact enveloping dimensions and geometry of the dummy hand, as shown in the figure. This mechanical synergy implementation was planned for installation in forearm since it is unlikely such a complicated transmission system could be fully embedded in palm. All the outputs from this implementation shall be connected to the prosthetic hand to drive the finger joints. Decisions on i) how to realize the mechanical synergy implementation, ii) how to actuate finger joints, and iii) how to connect synergy outputs to finger joint axes, should be made consistent and compatible with each other.

It was attempted to use planetary gears to realize the mechanical implementation of postural synergies since it might provide better accuracy than differential pulleys and might be easier to fabricate than designing a differential hydraulic system. Then the synergy outputs will be rotations (instead of being translations as in the other two options). The most direct way to connect these synergy outputs to the finger joint axes is to use flexible shafts. What’s more, rotations of these flexible axes would not be affected by a possible presence of wrist motions. Using worm gears and gears, rotations of these flexible shafts drive all the finger joints.

Two versions of the prosthetic hand were designed, fabricated and assembled. Version 1 was abandoned due to the bulky appearance and the flimsy construction. Version 2 has a modified internal structure with more precisely manufactured components. The hand can not only rotate the two rehabilitation training balls but also perform various grasps.

Version 1 Research Prosthesis2012 3.jpg
Version 2 Research Prosthesis2012 4.jpg
Research Prosthesis2012 5.gif
Research Prosthesis2012 6.jpg
Personal tools