Finding wavelet coefficients by applying sliding window operation on multivariate signal.

Hey friends!

I am trying to preprocess multivariate electrical signals (3 voltages, 3 currents) in timeseries form by applying sliding window operation. I know about the functions and how to use them. But I need to find wavelet coeffients of subsignals of complete signal (sliding window operation) and oraganize those coefficients in a sequence. Can anyone let me know about any built-in function or script for this type of customized signal processing?

Thanking in advance.

Answer 1

[Aastha]

Hi @Rab Nawaz,

As I understand, you want to apply a sliding window operation over multivariate electrical signals (3 voltages and 3 currents), compute wavelet coefficients for each subsignal in the windowed segments, and organize those coefficients sequentially. You may refer to the steps mentioned below to do so:

1. Use the "buffer" function to segment your signal into overlapping or non-overlapping windows:

% Example for one signal channel
window_length = 128;
overlap = 64;
segmented_signal = buffer(signal, window_length, overlap, 'nodelay');

You can apply this to each of the six channels separately or stack them if needed.

Kindly refer to the MathWorks documentation for any queries on "buffer" function:

https://www.mathworks.com/help/signal/ref/buffer.html

2. Use the discrete wavelet transform "dwt" function in MATLAB to compute the wavelet coefficients for each windowed segment:

% Compute wavelet coefficients for each segment
num_segments = size(segmented_signal, 2);
coeffs = cell(1, num_segments);
for k = 1:num_segments
    coeffs{k} = modwt(segmented_signal(:,k), 'db4', 4); % or use dwt
end

You can specify the wavelet type as an input argument to the "dwt" function and number of decomposition levels as per the requirements.

For more information on the "dwt" function and the types of wavelets available you may refer to the MathWorks documentation links mentioned below:

• dwt: https://www.mathworks.com/help/wavelet/ref/dwt.html
• Types of wavelets: https://www.mathworks.com/help/wavelet/ref/wavemngr.html

3. To organize the coefficients as a sequence, you can use the "cell2mat" function in MATLAB to convert the cell-array containing your wavelet coefficients to a matrix of size num_segments x coeff_length using the below MATLAB code snippet:

% Concatenate coefficients into a sequence
wavelet_sequence = cell2mat(coeffs'); % size: (num_segments x coeff_length)

For more information on the "cell2mat" function, you may refer to the MathWorks documentation:

https://www.mathworks.com/help/matlab/ref/cell2mat.html

I hope this helps!