About two or three years ago I had an idea. I was working in the University of Canterbury (NZ) Library, and thinking of writing cuneiform directly to the computer, using a tablet.
My idea was to use layers of electrotextiles
formatting link
thin rubber sheets between the layers to give the same "depth" as skin. The electrotextile layers would be at 45 degrees to each other to minimize the chance of artifacts.
I then realized that this would offer a possibility for creating a virtual or artificial skin that would allow robots to gauge the amount of pressure necessary to grasp things. I got in touch with the local representatives - Wooltech I think they might've been called, or some such name, based in the township of Lincoln - but they never got back to me and I didn't have the time to track them down.
I raised the idea with the MIT robotics prof guy, but he hasn't replied.
So, am I a nutcase, or is it my bad breath? Ie, will my idea work?
I can't view the site 'cause I haven't got flash and it kicks me out :o( Is it a pressure pad like burglars are supposed to step on to set off the alarm?
I find that people tend to develop their own ideas rather than other peoples, there's more money to be had that way and life's to short.
What you are describing is a Force Sensing Resistor (FSR). These have conductive trace patterns on one surface and a conducting type of rubbery substance on the facing material. Resistance is inversely proportional to pressure.
I had a similar concept some years ago. I envisioned a rubbery cloth probably made of a skinning foamed polyurethane coating with reinforcing fibres and a conductive filler (graphite?). The skin would contain a square or preferably hexagonal mesh of low power processors as a raw die or a minimally encapsulated die. The mesh is highly-connected for redundancy, i.e. each cell is connected to all adjacent cells. Each cell also has a data/sensor wire to each adjacent cell, and the data wire is *uninsulated*. Each data connection needs a tri-state transceiver and is multiplexed to an A/D converter.
By cooperation between the nodes, an adjacent node drives the data wire high and low while the local node detects the levels, thereby measuring the resistivity of the medium and hence the surface pressure, communicating its sense of touch by messages that ripple throughout the matrix. Every node would need some calibrating adjustment to allow for manufacturing variation, but this could be built in and performed during manufacturing. The redundancy means that if some wires are broken or some nodes dead, the power and data signals remain intact.
If implemented in low speed CMOS, the power requirements would be so low that power and data connections from the outside world to the matrix as a whole could perhaps be contact-free, by induction directly into the power wiring (remember that being redundant, it forms loops).
The idea here is to fabricate a skin that can be bought by the yard for under $1/cm^3, and can be cut with scissors, glued, sewn, stretched or whatever, into almost any shape, and form a tactile skin over any surface, which can be pulled over any form containing a power/data transciever and "just work".
The idea's not bad, though it wasn't quite the idea I was working with. Still, it's worth working on, working out the possibilities and one fine day we may have something that works, as well as sounding new and useful. ;)
Electrotextiles are a range of textiles with fine wires interwoven among them. Their conductivity changes with pressure, and this has been used in several trial products, to the best of my recolleciton.
My idea is simply to use that conductivity change - in conjunction with rubber "buffer" layers, or maybe a proper "rubberization" of the textile - to indicate small changes in pressure, sufficient to train a robot to grasp objects in a humanly-satisfactory manner. I'm talking about gearing robot grip to the same sort of sliding scale as we humans use.
And this seems to me to be one way of doing it, at a relatively low cost. Textiles tend to the lower end of the consumables marketplace, which is why there is a "fashion" industry geared towards making us pay more for less.
The idea of changing resistance with pressure can be very responsive, they used carbon granules compressed by a diaphragm in 1920's telephone mouth pieces.
There have been a number of these developed over the last 20 years or so. A "skin" is not the common aim, but rather tactile feedback in robot end-effectors. There is far more money in industrial process applications than making a version of Data from Star Trek. At least for now.
As many of these ideas end up being patented, people have tried all sorts of alternatives, including resistive, capacitive (tiny air bubbles separating a grid and a single sheet of metalized plastic), fiber optics, ultrasonic sound, changes in heat, and literally dozens of other approaches.
Some you can make yourself in basic prototype form. A tape product from
3M exhibits conductivity in the Z-axis only, and the conductivity changes slightly with alterations in pressure, especially when multiple layers are used. Though the output is not linear, one could place a grounded thin metal foil over this tape. On the bottom is an array of conductors, like a printed circuit board. A home-made version would not be flexible, but it's quite possible to make flexible circuit boards.
My advise is to thoroughly check the literature. Some of this has been written up in magazines like Popular Science, and a good deal in Science News. Online editions of both are available, but a good library is better. Though you are in New Zealand, the US patent office is also a useful reference point. You can search through the patents by keywords. Go to
formatting link
Most patents will reference (as differentiations of prior art) other patents as well as published articles.
[Zagan] Tux, have you considered using variable capacitance for such an application? If you place two conductive sheets (e.g., aluminum foil) layered over a non-conductive foam material (thus making a capacitor), you can detemine if pressure is applied by measuring the time-constant of the capacitor in series with a resistor.
PolyTech Forum website is not affiliated with any of the manufacturers or service providers discussed here.
All logos and trade names are the property of their respective owners.