Criteria for the final SINR, CQI computation in 5G NR CSI Reporting

In 5G toolbox, specifically NR Downlink CSI Reporting Example, SINR is being calculated for all layers. I have a question, for a two-layer scenario, if SINR values vary considerably, what should be the criteria for the final SINR, CQI computation? The lowest SINR, the mean or something between? In other words, which one of these criteria should we choose, as each comes with pros and cons.

1. Go with lowest SINR.
2. Go with higher SINR and rank1.
3. Go between lower SINR and average SINR.

Thanks

Answer 1

[Umar]

Hi @Jake,

Addressing your query regarding, “In 5G toolbox, specifically NR Downlink CSI Reporting Example, SINR is being calculated for all layers. I have a question, for a two-layer scenario, if SINR values vary considerably, what should be the criteria for the final SINR, CQI computation? The lowest SINR, the mean or something between? In other words, which one of these criteria should we choose, as each comes with pros and cons. 1. Go with lowest SINR. 2. Go with higher SINR and rank1. 3. Go between lower SINR and average SINR. “

Please see my response to your comments below.

Given the considerations outlined above, a balanced approach that combines lower and average SINR is often recommended for computing CQI. This method provides a compromise that can enhance overall network performance while maintaining reliability.

To illustrate this approach using MATLAB, we can simulate SINR values and compute the final SINR as a weighted combination:

% Example SINR values for two layers

SINR_layer1 = 10; % in dB  % Define the SINR for layer 1 in decibels

SINR_layer2 = 20; % in dB  % Define the SINR for layer 2 in decibels

% Convert dB to linear scale

% Convert SINR of layer 1 from dB to linear scale

SINR_layer1_linear = 10^(SINR_layer1/10);  

% Convert SINR of layer 2 from dB to linear scale

SINR_layer2_linear = 10^(SINR_layer2/10);  

% Compute final SINR

% Find the minimum SINR value between the two layers

lowest_SINR = min(SINR_layer1_linear, SINR_layer2_linear);

   % Calculate the average SINR of the two layers

    average_SINR = mean([SINR_layer1_linear, SINR_layer2_linear]); 

    %Weighted combination (you can adjust weights)

    weight_lowest = 0.4;  % Assign a weight to the lowest SINR

    weight_average = 0.6;  % Assign a weight to the average SINR

    % Compute the final SINR using the weighted combination

    final_SINR = (weight_lowest * lowest_SINR + weight_average * average_SINR) 

    /(weight_lowest + weight_average);  

    % Convert back to dB

     % Convert the final SINR from linear scale back to dB

     final_SINR_dB = 10 * log10(final_SINR);  

     % Display results

     % Print the final SINR value in dB format

     fprintf('Final SINR (dB): %.2f\n', final_SINR_dB);  

Please see attached.

Also, by balancing between different SINRs, operators can enhance user experience while maintaining efficient network performance. Please bear in mind that in real-world implementations, one should also consider additional factors such as user mobility, interference from neighboring cells, and overall network load when deciding on CQI calculation methods. Hope this helps. Please let me know if you have any further questions.