Forums

adaptatif filter lms

Started by abir hadrich February 20, 2008
hello,

im new to the topic of adaptive filters and its implementation in matlab. im doing a project on 'filtering an ECG foetus'. we are carrying out all our analysis on an ECG wave with only a powerline interference (60hz).

the signal available to us is an ECG wave with a noise component of 60hz, sampled at 360hz..
can you help me implement an adaptive filter using LMS algorithm for this signal in matlab... i am not clear with regards to the value of mu(step size), filter order, necessary inputs to be given to filter etc... can someone please send me a SAMPLE CODE.

thanks
hello,
>
>im new to the topic of adaptive filters and its implementation in matlab. im doing a project on 'filtering an ECG foetus'. we are carrying out all our analysis on an ECG wave with only a powerline interference (60hz).
>
>the signal available to us is an ECG wave with a noise component of 60hz, sampled at 360hz..
>can you help me implement an adaptive filter using LMS algorithm for this signal in matlab... i am not clear with regards to the value of mu(step size), filter order, necessary inputs to be given to filter etc... can someone please send me a SAMPLE CODE.
>
>thanks
Hello abir,

please have a look at my website:
www.stefantwieg.de.vu
there you will find c# / matlab source code for adaptive filter implementation with NLMS. The sample is for system identification tasks but you can easily match it to your application. You only have to change input/desired signal. After generating a reference noise signal which is correlated to the interference you could use this as input x(n). The desired d(n) signal is the original noisy signal (ECG). The adaptive filter will calculate the weights with gradient descent method so that all correlating signals are remodeled and error is increasing.

regards
stefan