Progress Report 7 (3/20/19)
Previous Goals
Last week, we sought to test and finalize the acquisition and preprocessing code, troubleshooting the infinite loop error. We also strived to implement and test the EMG differential and buffering code and finish SNR analysis (determine a primary subject and find the ideal electrode placement). In terms of hardware, we wanted to incorporate LED shields to indicate the amplitude of muscle contraction for the subject. We wanted to research a potential sliding setpoint to more effectively implement in the PID software platform and utilize a phantom signal. Lastly, we intended to test the Parallax servo with an Arduino UNO to better understand how to integrate it with the PID.
Work Accomplished
Since the last progress report, our team performed recalculations of the SNR data, recollecting data from some subjects and correcting other SNR calculations. We also edited the CAD design for the hardware casing, including a curvature so that the case will now fit the arm instead of gapping on the sides and designing a lid to fit that curvature so that the hardware will now be portable. The design was also edited to fit the actual volume of the hardware components.
LED shields were implemented with the myoware sensors which allowed for relative simplifications in the circuitry. Because of some issues with headers, there was damage to a single myoware sensor and an LED shield. Trouble shooting was performed on both of these elements, but no solution could be found. Another shield and sensor will be purchased. A servo motor was tested with a pre-built arduino library to observe its behavior given various input forms. In order to better integrate our algorithm system into the required outputs for the servo motor, further work needs to be done in this area. Additionally, we have also secured a contact at Ottobock through whom we will be able to send the arm for repairs.
We continued our software development and changed the communication scheme for our algorithm to account for our new output requirements. Our main software efforts have been in the calibration, smoothing and rectifying of the EMG waves, and the buffer array definition in the arduino platform. A meeting with Dr. Goldfarb helped guide efforts in these areas and confirm the validity of our array approach that will be implemented in the near future. We switched from a moving average calculation of EMG signals to the calculation of a Tustin approximation for better compatibility with arduino data handling. The two efforts were compared using previously collected data sets in Matlab.
We continued to meet with our advisor and received feedback from her this week regarding the translation between the demonstration-based work of the servo motor and the ultimate implementation in the prosthesis. We are continuing to investigate how best to accomplish a smooth transition between these two forms and have accounted for that in our current efforts.
Work Backlog/What Went Wrong
- One of the Myoware sensors/LED shields is not performing as well as expected
- Does not show values to be as high as the other sensor
- Headers broke off when trying to implement LED shields which caused damage
- Difficulty with implementing sourced code
- Reconfigured plan of action
- Used Arduino’s stock “Servo” code but need to switch to a method which allows for modification of pulse widths more explicitly
- Still need to send in Ottobock arm for repairs
Plans for Next Week/How to Accomplish
- Continue to work on algorithm
- Implement the buffer management system
- Begin implementing differential algorithm
- Continue to work on Pulse Width Modulation for servo motor
- Determine what pulse widths map to what angles
- Determine if servo automatically goes back to a “set point” with no pulses
- Send arm to Ottobock for diagnosis
- Our contact said someone will be in touch soon with an address
- We will then ship it to that address
- We need to order another LED shield to replace the one that has stopped working
- Replace LED shield and configure new form for attachment, whereby strain will be reduced on headers
- Finalize relationship between hand angle and EMG input for gesture transformation