Weekly Meeting 9 (2/11/19)

Posted by Rebecca Jones on Monday, February 11, 2019 in Notebook.

Prosthetic hand update
- They say that we need an account number
- Emily will email her contact and cc Shaan
- If this doesn’t work, it might be best to show proof of contact with a servo motor
  - Disclaimer: it may be more work
    - Would have to recreate current control scheme
  - Should start finding a motor
Myoware
- Have an output
- Have not tested on phantom tester, but will begin that soon
- Will set up power source this week
PID
- Emily will look through document Allyson sent her
- Currently coded in MATLAB
- Looking into compatibility
- Integral control will be lowest level
- Proportional control will have the highest
- Coding environment:
  - Thinking of using Arduino from MATLAB
  - Matlab can deploy on arduino, but has to be connected to computer
  - If you make a control schema in simulink, you can deploy it onto the arduino and have it be an isolated, stand-alone system
  - Emily: Shouldn’t waste too much time trying to make that work
  - Make sure we comment the code well — most important
- Feature classification:
  - Current plan of action: KNN classifier
    - Groups into different categories based on cut-off
  - Should think about efficiency
  - Someone in their lab could help us with this
- Embedded systems constraints:
  - Biggest: computation time
    - Function of sampling rate
    - How frequently you’re asking it to cycle
      - Has to compute many things within loop
  - Memory management
    - Size of program and local data storage (onboard memory)
- Stimulation parameters:
  - Windowed
    - Set by simulation frequency
    - 50 ms should be fine — could even go slightly slower
      - Don’t want noticeable lag
        
        Notice at about 100-150 ms
EMG
- Will try to work on training sets and get a better idea of how it works