Indexing in 4D array

I am writing a program where I have 15 Eb/No values for each Eb/No there are 100 packets to be transmitted from Transmitter to reciever. Each packet can have multiple OFDM symbols and each OFDM symbol can have 132 subcarriers.

So my array looks like (EbNo,no_packets,OFDM symbols,no_subcarriers)

for each group of subcarriers I have to perform certain operations. I have to take the average of the power present on each subcarrier. I also have to perform some other functions.

Given the no of symbols are 2 , the average of matrix should be 2*15*100 matrix. But I am stuck in indexing . I have tried many possibilities but my avaerage isnt right.

Also I am passing values into other functions. So should I mention the indexing or not.

for EbN0id = 1:length(EbN0dB_vec)
    for np=1:Nof_TxPackets 
        %------------coding and modulation 
        [x_time,dataIn,dataMod] = LDPC_Coding_and_Modulation_Power(Nof_BitsPerPacket, poly, bgn, Length_dataEnc, rv, ModOrder(m), nLayers, Nof_OFDMSymbols, Bit_x_symbol(m), Nof_Subcarriers, Nfft, Ncp,optimal_power);             
        %-----------Add Multipath Channel
        if Multipath_Flag
            [y_time,H_k] = AddMultipathChannel(DelayProfile_samples, Fading(fi+1:np*Nof_OFDMSymbols,:), Nfft, Ncp, fi, Ts_OFDM, Nof_OFDMSymbols, x_time, Nof_Subcarriers);             
        else
            H_k = ones(Nof_OFDMSymbols,Nof_Subcarriers);
            y_time = x_time; 
        end       
        
        Average_power(:,EbN0id,np)=(sum(optimal_power(:,:,EbN0id,np)))/Nof_Subcarriers
        
        gamma(:,:,EbN0id,np)=Gamma_SNR(variance_of_noise, optimal_power(:,:,EbN0id,np),aver_power(:,EbN0id,np),H_k(:,:,EbN0id,np))  ; 

Answer 1

[Pratik]

Hi Shama,

In my understanding, you have a 4D array “optimal_power” with the dimensions: “EbNo, no_packets, no_OFDM_symbols, no_subcarriers”. You are attempting to calculate the average power for each packet, each ‘Eb/No’ value and each ‘OFDM’ symbol where the result should have the dimensions: “no_OFDM_symbols, EbNo, no_packets”.

Taking the average to get average power involves summing over the last dimension of the array which is “no_subcarriers”, then dividing this sum by the number of subcarriers and finally using the “permute” function to get the array in the expected dimension.

The following code snippet demonstrates the use of “permute” function and how the average can be calculated:

% Calculate sum over no_subcarriers dimension 
sum_power = sum(optimal_power, 4); 
% Calculate average power 
Average_power_default = sum_power / Nof_Subcarriers; 
% Permute the dimensions 
Average_power = permute(Average_power_default, [3 1 2]); 

“Average_power_default” will be a 3D array with dimensions: “EbNo, no_packets, no_OFDM_symbols”, which represents the average power for each ‘Eb/No’ value, each packet, and each ‘OFDM’ symbol. "Average_power" represents the array with expected dimension.

While using the “Gamma_SNR” function, you should pass the entire 4D array “optimal_power” and let the function handle the indexing itself. If the need is to pass only a specific part of the array the”:” operator can be used to select all elements belonging to a specific dimension.

Please refer to the following MathWorks documentation link for more information on “permute” function:

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

Please refer to the following MathWorks documentation link for more information on array indexing:

Array Indexing - MATLAB & Simulink - MathWorks India

Hope this helps!