Whereas one may try to optimize grip quality by increasing finger control and grip geometry sophistication with over-underactuation, increased amount and thus weight of electronics and batteries, increased risk of failure and breakage due to increasingly finely structured constructions of endless sophistication, grip quality itself is a different subject. Not altogether, but somewhat. Due to that, grip optimization also is a somewhat different subject.
As I already established in a previous article, a non-friction slippery grip will require something close to seven fingers in three dimensional space. Anything less in terms of fingers will require increasing amounts of friction.
In an extreme view, this allows to extreme options:
- A fully controlled seven finger gripper that may feature metal, glass, hard plastics or similarly undeformable surfaces. Control then requires full grip contact to the object.
- Or a two finger gripper that contains maximally adaptive surfaces that deform and impose as much friction as ever necessary on the object that is to be gripped.
With a minimal control - cable control, pull to open, let go to close - minimal forces to open the gripper are a must have because of compression issues (see separate post not yet written). This constraint also reduces the available average or maximal closing force of a voluntary opening device.
With just two prosthetic gripper fingers, this places burden on harness design, cable control and grip maximization onto the non-slippery surfaces. I already massively improved cable design (one down), we are currently building a new shoulder harness (two down, work in progress) and so the last element standing is grip. So, increasing friction remains the ultimate issue to two-finger gripper issues for upper extremity prosthetics.
With this, almost all is said. Obviously, soft surfaces deform better. Therefore, soft surfaces are what we are after. Also, soft surfaces decay faster, fall apart, get damaged, rip apart, and have to be replaced. Thus, availability and ease of changing materials is the next logical requirement.
For grippers such as the V2P Prehensor, glove fingers or tape options come into considerations. For prosthetic hands, only hands that can be equipped with affordable gloves are real options - all gloves that cost over a few hundred bucks are already a non-option to begin with. The Becker hand is cool, it works with just about any stock work glove - fabric, leather, PVC, latex, nitrile covered - available. The Otto Bock system hand is a lot more difficult to equip with stock gloves so it is not so recommendable from that point of view.
With a Becker hand that contains an adaptive grip, surface may be somewhat deformable yet somewhat slippery as more geometric constraints for a gripped object reduces the grip burden on the surface. Currently I am wearing a PVC glove and despite its somewhat slippery nature, it works very well.
With a V2P Prehensor, however, surface grip friction is everything. Recently in a discussion with Alexandro Hernandez Arieta, currently at the AI Lab of the Institute of Informatics at the University of Zuerich and Lijin Aryananda, currently Managing Director of the NCCR Robotics in Switzerland, a newly developed type of product termed "Gecko Skin" was mentioned as Lijin pointed out that most of my prosthetic needs would probably not be found in the field of mechanics or robotics but material science. Material science - true. Gecko Skin - great tip.
So before the sun went down and came back up, I had already ordered a batch of Arundel Gecko Handlebar Tape.
Gecko Skin 1 - Arundel (R) Gecko Handlebar Tape
Works nicely. No grip force measurements out just yet.