05-04-2011 11:20 PM
Hi beirut,I am finding it somewhat difficult in designing a classifier.Can u post some helping models or so.Any help can u provide to move further.Thank u.
05-05-2011 01:10 AM
Am working with neural network, so i will give you the process how to make the model.
If you have never studied neural network you have to read about it.
1- The neural network uses a database in order to form its model, and at the end of the process it is formulated in a mathmatical formula so can be used and programed within your application.
Take this example:
Suppose you have air conditioner and you want it to work at low speed if the temperature is between 15-20, medium speed 20-25, and high speed 25-...
so the input of your system is the temperature.
collect the database of different temperatures you have.
2-Since you are designing the system you know for each input what is the desired output you want for this specific input. So match each input to its output.
3- now you have array of inputs which have mapping to an array of outputs.
These 2 arrays are used by Matlab or Labview in order to give you the model of the neural network.
If you don't need to use softwares to design the neural network you can do it by hand but you have to study the theory.
07-13-2011 01:38 PM
Hello..
I would like to share the prototype of my project while testing it in this link
07-13-2011 04:10 PM
Looking good! Looks like you have nice stable control going with more than one degree of freedom. Nice job! Is the code done in LabVIEW?
07-13-2011 04:18 PM
Thank you steve
No the code is implemented on digital signal processor so it was written in c language, but i used ni elvis osclliscope and dynamic signal analyzer in order to monitor the signal and to see the frequency response of the signal.
10-18-2011 11:55 PM
Greetings!
In beirut's post of 28 December 2010, he said:
The Filter i designed is based on butterworth filter design, it is formed of 2 filters, the first one is low pass, and the other is high pass filter,
to make band pass filter f1=20Hz and f2=500Hz
Can someone explain why these two filters are required? Could either or both be eliminated?
10-19-2011 10:10 AM
Hi ignoramus,
when the emg signal is acquired, it have your frequency spectrum between thoses values (20 - 500 hz).
Others frequencies values (below 20hz or above 500hz) are just noise.......
You should not eliminate the filters, but, instead using a low pass and a high pass filter, you can just use a band pass filter
Hope I'd helped.
10-19-2011 02:10 PM
EMG signal frequency component ranges from 20Hz to 500Hz, so in order to acquire the components that are related to emg signal you have to eleminate frequency component that are not in the range of emg signal. But the big problem that you will still face is the noise generated from 50-60Hz power ac lines. Notch filters may be used to eleminate this frequency but this will cause to eleminate another components since there is no ideal filter, so i didn't use it in my design.
10-21-2011 06:59 AM
Thanks to beirut and RamonCampos for answering my questions.
I would like to buy an amplifier with filters to use with an A/D converter.
If anyone can supply one, please contact me direct at cjsdesilva@gmail.com
12-19-2011 10:41 AM
Hi,
I am doing a similar project for my Btech final year. This is what I intented to do- I want to perform 5 different fuctions using my artificial hand (robotic arm)- 1) Elbow flexion 2) Elbow extension 3) wrist extension 4)wrist flexion 5) finger extensions. I thought of taking EMG signals from the triceps and biceps while performing these movements and then feed these signals to the MATLAB unit- where I will be training the system, such that when a particular pattern comes, a particular movement of the robotic arm will be performed.
But then I am struck with the initial phase of the project. I seriously need big help with the .dat files of the signals told above. Can smeone please help me by giving the data files of the above movements taken from the biceps and triceps??
It would also be great if you could give your expert advices for the project.