Controlling a prosthetic hand

I took an EMG signal of 17 hand gestures from 43 person separately and now I have the dataset of these signals. After doing segmentation, DFTI, Feature-Extraction and classification how can I identify the the movement to the servo ?

[Vaibhaw Kumar]

Did you get it?

Answer 1

[Ayush]

Hi @Ahmed,

I understand you need to find methods to identify the servo movements.

Tracking Servo movements can be achieved using a two-stage process:

1. Classification: you can use a suitable machine learning algorithm to classify the gesture. Some of them include:

• Support vector machine (SVM): You can read more about it here: https://www.mathworks.com/help/stats/incrementaloneclasssvm.fit.html
• Random forests: You can read more about it here: https://www.mathworks.com/matlabcentral/fileexchange/31036-random-forest
• Neural networks: You can read more about it here: https://www.mathworks.com/help/deeplearning/ref/dlnetwork.html

2. Mapping: Once you classify the gestures, you will need to map the gesture classes to servo commands. For example:

• Gesture 1 -> Move servo to Position A
• Gesture 2 -> Move servo to Position B

And so on…

3. Implement a control loop in your software (eg using Arduino) that listens for classification results and sends commands to the servo motor accordingly.

You can read more about Arduino support with MATLAB here: https://www.mathworks.com/help/matlab/supportpkg/arduino.html

Hope this helps!