Ok, purchased more spring scales that work correctly, and have metric measurements on them. I am trying to decide what range of forces are considered to be 'low', 'med', and 'high'. So far, low forces that would still be detectable seem to be somewhere around the 1-2 N range. Medium is somewhere between 3-6 N, and high is probably greater than 7 N or so. The range is not evenly spaced or linear, in my opinion.

Ok- so Li and I started testing ourselves. We've tried several variations of tests. Using some foam-tape and fishing line, we stuck the tape on the skin and pulled at the string using a spring scale to keep track of the forces administered. We devised a list of 18 numbers, each either a 1, 2, or 3, where a 1 signified a low force, 2 = medium force, and 3 = high force.

We first tested Li's forearm and back of hand. I gave Li the initial low and high forces to set some sort of range. Then we used a sequence of 18 different forces, either low, med, or high to test Li. At first, I simply applied the force quickly, then asked Li what force it was.

He seemed to be able to distinguish the forces fairly easily, although for some reason, in every test, he got the first one incorrect. He said that he learned what the forces felt like, so it was easier as the testing progressed to distinguish them. However, we felt that it was unlikely that a user would experience all these varying forces in a short amount of time. It seems that Li found it easier as time progressed because he could remember the forces given previously and use them to compare the force was currently applied. Also, low forces were easy to detect, but medium and high forces were harder to distinguish. One factor that helped the user to feel the high force was that the high force caused some small amount of pain, so Li said he knew it was the high force if it hurt.

We decided to apply the forces and keep them there, to simulate the DC response better until the user responded with what force was applied.

Then it was my turn- I noticed the same things that Li noticed. Low forces were very easy to detect. However, during one of the tests, Li put his hand on my arm, while he pulled on the spring scale, and I noticed that it was significantly harder to feel the force. So, we decided to test how well I could distinguish the skin stretch forces with Li?s hand on my arm. We thought it would be more realistic, because a haptic device would most likely be strapped onto a user?s arm. Also, because when we did the tests with the poking device, I noticed a similar problem, in that the cloth wrapped around my arm caused more sensation than the actual poking.

So, with Li's hand touching my arm while he applied various forces, it was a little harder to distinguish the forces. Low forces weren't even felt at all. The only reason I knew that the force was applied was because Li would ask for me to tell him which type of force he was applying. Obviously, because I could not feel anything, I figured it was the low force. Distinguishing between the medium and high was more difficult, but I was still able to get most of the forces correct (34/36). Without Li's hand on my arm, I was able to correctly determine what type of force was applied (36/36). I noticed that the difference in forces was easier to feel on the back of my hand. However, the amount of pain felt when a high force was applied, was more pronounced than the pain felt on the forearm. So each person is going to have a varying amount of skin stretch and it may be more painful for some than others.

Also, we decided to re-test Li using the same, hand on arm/dc response force application. It was interesting to note that earlier, when we were testing ourselves, no one else was really in the lab and it was quiet. However, the second time Li was being tested on, more people came back into the lab and began crowding around and commenting/laughing/talking. And then Li began to incorrectly describe some forces. He said it was distracting, having the members of the lab crowding around, and it made it more difficult for him to concentrate on the forces being applied.

Low force is almost TOO low. Especially when the hand was also on the arm of the skin being stretched.

Stretching on the back of the neck using the method we are using is difficult because when stretching the skin, there is a normal force that pulls the tape off the neck easily. It's harder to get a shear/tangential force on the neck than the arm. The shoulder is also difficult to get to simply because of clothing.

It would also be interesting to complete a JND test for the forces...it's most likely not linear...

Conclusions (sort of, obviously it's not very conclusive with only two test subjects) from our preliminary testing today? It's possible to distinguish generic, low, med, high forces, from one another, and I think further pilot testing will show that most other people can distinguish the different forces, but I have a feeling that combined with the feeling of the haptic device attached to the body, and the problems of various outside distractions, and the ability (or lack of) of humans to remember the previous forces applied, that this will become more of a binary (on/off) type of haptic application. Either you feel it or you don't.