Progress Report 5 (2/13/19)
Previous Goals
Our prior goals included writing a protocol for EMG studies with myoware sensors (such as how to test and where to place electrodes). We also strived to investigate EMG signal interpretation and analysis, reading existing literature to map expectations of movement. We further investigated PID behavior with manipulated variables creating sine and step functions for previously-written MATLAB code. Additionally, we wanted to get set up on Github for code communication. Finally, we set out to get in contact with our sponsor’s contact at Ottobock to determine next steps.
Work Accomplished
We established a protocol to acquire EMG signals. This included the process of actually obtaining the data and certain movements that we will use. To do this, we will extend and flex our wrist to activate the flexor digitorum and extensor digitorum. This also involves a plan to test multiple participants from our team to determine which team member we will use as our phantom test subject. We were able to establish a GitHub repository for code sharing between the team members working on algorithm development to prevent overlapping efforts. Research was done into the coding platform that should be used. Due to a disconnect between the compiling software that would allow for Matlab use and the development board, all further coding was done and will continue to be done in Arduino IDE. Investigations were performed into the PID behavior and the existing code was tested further using new input functions that were designed. Using varied sine wave and step function observations, it was found that a fraction of constant control Proportional: Integral:Differential of 1: 0.008: 0.5 was optimal for control of an EMG output towards our set point of a relaxed hand position. Further research was done into the feasibility of a machine learning processing approach in gesture mapping within our program. Our hardware was received and the myoware sensors were attached on independent analog outputs. The teensy adaptations for arduino software use were installed and the initial analog outputs from the sensors were viewed as serial plots. Ottobock communication was continued and the identification of the need for a prosthetists was reaffirmed, further work is being done with our advisor to identify the best alternative route. We anticipate either sending the prosthesis to Ottobock for repair or moving forward with a different proof of concept/validation method.
Work Backlog/What Went Wrong
- Ottobock communication continued without advancement due to requirements that academic research teams cannot fulfill
- Need for prosthetist cannot be met
- Efforts for resolution will be continued
- Sponsor will provide information for Ottobock contact that the lab at Case Western has used
- Team is still investigating the possibility of using a servo motor instead of obtaining a new hand from the Case Western lab
- Simulink was incompatible with Teensy, preventing us from performing more coding in MATLAB
- Initial EMG testing using Labview was determined to be extraneous
- Arduino software required for EMG data acquisition was only able to work on one team member’s computer
Plans for Next Week/How to Accomplish
- Continue Hardware Addition and Construction
- Research benefit of direct communication between myoware sensors and attempt connections
- Attach battery power to remove need for computer hookup
- PID implementation in Arduino platform
- Translate existing matlab code to consider the sliding window, dynamic platform in arduino language
- EMG testing
- Begin acquiring EMG signals from team members
- Begin determination of which team member to use for phantom testing based on clarity, accuracy, and amplitude of signal
- Become familiar with initial signal output
- Determine whether ground at wrist or elbow will provide cleanest signal
- Further investigate equations for signal processing
- Investigate myoware processing capabilities
- Continue literature search
- Begin acquiring EMG signals from team members
- Embedded Systems Development
- Hold phone meeting with embedded systems subject expert at Case Western
- Research best practices for sliding window and memory management
- Begin developing basic algorithms and controllers for Arduino
- Contact sponsor’s point of contact at Ottobock to determine how to send the prosthesis for servicing.
- Depending on how this goes, we will either move forward with getting the prosthesis serviced or find another device to test actuation