EMG Signals Drive Servo

This simple project is what I really want to do before. I used Biopac system to amplify my EMG signals and acquired these signals by NI USB-6009.

Originally, I want to use serial communication to send the control signals to arduino, but I don't know how to send the unsigned long integer in labview instead of string. As the result of that, I used USB-6009 to generate analog output to arduino.

Every hardware thing is ready, next is to write the program. As I mentioned before, I used labview to acquire EMG signals, so I just acquired the signals, calculated by RMS, and transformed the number into the range 0-5 to have the USB-6009 generate analog output voltage between 0 and 5.

The arduino program is also very simple. It just uses analogInput to receive the voltage from USB-6009 and controls the servo by the voltage value.

However, as you can see, the control is not good. I think this is the key point that I need to establish a good control strategy in the future.

EMG Median Frequency

Now we try to calculate the median frequency of this raw EMG data:

1. We need to use fast fourier transform.

2. Integrate the half of the data, because the other half is image of the real data.

3. Find the median value of this Integrated data.

    Here is my matlab function:

EMG RMS

Now we want to calculate the root mean square of this raw EMG data:

Here is my code:

window means how many points do we want to calculate one RMS.

over means how many points do we need to overlap.

You can also find the example in my matlab blog: (Chinese Version)

http://labmatpt.blogspot.com/2006/12/moving-averagerms.html

EMG Filter

Here is a Matlab code for high pass, low pass and band pass filter:

Take low pass filter for example:

I wrote the code by modifying the "butterworth" filter function.

We can switch to high pass, band pass or band stop by changing the string 'low'

Research

EMG
    
    Time Domain
           EMG Onset
           EMG Filter
           EMG RMS
      
    Frequency Domain
        EMG Median Frequency


Sonography
    My thesis
        Tendon Viscoelastic Properties
    ISB Conference
        The Correlation Between fascicle Curvature and Muscle Hardness
    Recent Study
        Elastography
        
Publications
Wu YK, Lien YH, Lin KH, Shih TF, Wang TG, Wang HK.  Relationships between three potentiation effects of plyometric training and performance. Scandinavian Journal of Medicine & Science in Sports: in press  (SCI) PDF

Wang HK, Wu YK, Lin KH, Shih TF.  Noninvasive analysis of fascicle curvature and mechanical hardness in calf muscle during contraction and relaxation. Manual Therapy 2009; 14(3): 264-269  (SCI) PDF

Chiu CM, Wu YK, Lin KH; Shih TF; Lien YH; Wang HK. Effects of Plyometric Training on Eletromechanical Delay Human Plantarflexor Muscles. FJPT2008; 33(1): 34-40

EMG Onset

Here is an example to try to find the onset of EMG in electromechanical delay:

The definition of electromechanical delay here is the time delay between the M wave which

is stimulated by ES and the onset of force.

Finding force onset is pretty easy. I just found the point which is double standard deviation of

the baseline force.

However, finding M wave onset is a little bit difficult. I made two definition to find the onset 

point. One is the location of point should be in front of the location of Max peak and Min peak of M wave.

The other is the point should be zero after differential.  

EMG paremeters in My thesis

In my thesis, I analyzed EMG in time domain and frequency domain.

In time domain: the root mean square EMG

In frequency domain: Median frequency