The past week has been a successful one in terms of getting the hand to actually *do* something. I’ve gotten the hand to move a finger (specifically, the index finger), which is a pretty big step forward in the project overall. Check out the video below!
So, I was able to read the muscle and activate a servo based on the threshold reached. This is a bare bones version of the final control system, and it’s safe to say that this is the system we’ll be keeping. What happens next?
Hopefully in the coming week I can set up a meeting with the recipient of the hand to see exactly what muscles can be read. Since he doesn’t have a hand, I’m unsure of whether or not he’ll be able to produce a signal with his forearm muscles – we’ll just have to wait and see. Before the meeting, I’ll have the task of modifying the code for data collection, which should be easy to accomplish. After the testing, I’ll probably just put the raw data into Excel to make a graph, which I’ll then analyze. Alternatively, I may just take video of the tests, so that there will be a visual association with each spike in readings from the MSK.
On the note of testing, last week, I ordered four more muscle sensors, plus a bunch of other equipment. When it arrives, I’ll use the new equipment to read multiple muscles at the same time, and to build a “sensor sleeve” – that is, a sleeve that slides over the arm with reusable sensors built in. If the equipment arrives before the meeting, I may even be able to do a more thorough test to determine which muscles are usable. Then again, it is uncertain as to whether or not a meeting will occur – I have a short week this week due to my quickly approaching senior project.
While I did spend most of my time experimenting with control systems last week, I also did some experimentation with the motors and with our Form 1 stereolithographic 3D printer. I was able to get the Form 1 to work again (it has been out of use for the past few months due to a scratch in the silicon layer of the resin tank), and we may use it to print a few hand parts. As for motors, I’ve determined that servos will be utterly useless in the final version of the hand, as they require power to stay in position. Instead of servos, I’ll probably end up using small, normal DC motors with a worm gear setup. With such a setup, the fingers will be unable to be moved by something other than the motor itself, and they will not require power to hold a position. This system will probably require that I use a sensor to get an idea of where the finger is at any given time, but I’ll figure that out when I have to. Right now I’m focusing on controls.
As a final note, I looked into replacing the elastics on the hand with string. I’ll likely upload a video of this new system in action next week. Basically, instead of the elastics bringing any finger back to its starting position, it is the motor position that determines where the finger is (strings are used to open AND close the finger). The elastics that we’ve used have been unreliable, so this system should eliminate that issue.