Triangular-weighted moving average filter

Hi,

I`m looking for function or source code for Triangular-weighted moving average filter to apply it and make my data processing.

High-frequency noise shall be removed from the measured signals using a triangular-weighted moving average with a smoothing width of 100 ms.

can someone share the experience ?

[Jon]

Is this a homework problem? If so please see https://www.mathworks.com/matlabcentral/answers/8626-how-do-i-get-help-on-homework-questions-on-matlab-answers

[Mateusz Rzeszowski]

nope

Answer 1

[Image Analyst]

Use conv and set up your kernel to be a triangle shape.

[Dan]

Try:

tri_weights = triang('sz');

filtered_data = conv('data', tri_weights, "same") / sum(tri_weights);

Answer 2

[Meg Noah]

Here's an example of high frequency noise being removed with a triangular filter, if by width you mean the base of the triangle:

signal = repmat([zeros(1,500) ones(1,2000) zeros(1,500)],1,10);
time_ms = 0.1*(1:numel(signal));
dt_ms = time_ms(2)-time_ms(1);
filter = triang(round(100/dt_ms))/sum(triang(round(100/dt_ms)));
fprintf(1,'Sum of energy conserving filter should be 1 = %f\n', sum(filter(:)));

Sum of energy conserving filter should be 1 = 1.000000

fprintf(1,'Filter width = %d samples = %f ms',numel(filter),numel(filter)*dt_ms);

Filter width = 1000 samples = 100.000000 ms

smooth_signal = conv(signal,filter,'same');
plot(time_ms,signal,'b','DisplayName','Signal');
hold on
plot(time_ms,smooth_signal,'r','DisplayName','Smoothed Signal');
legend('location','best');
xlabel('Time [ms]');
ylabel('Signal');

●

You can also apply the convolution theorem to do it with Fourier transforms.