Clarification on Power Allocation in CLSM Feedback SINR Calculation

Vienna 5G System Level Simulator
1

Dear Vienna 5G Platform Developers,

I have a question regarding the SINR calculation in CLSM feedback, specifically about how signal and interference powers are computed in the current implementation.

In the current code:

% compute macroscopic, desired receive power
txPower = [obj.lqm.transmitter.transmitPower];
rxPower = sum(txPower(obj.lqm.desired) .* (obj.lqm.macroscopicFadingW(obj.lqm.desired))) ...
    /obj.transmissionParameters.resourceGrid.nRBFreq/nLayers;

% interference power
intRxPower = reshape(sum(obj.lqm.powerRB(iRBFreq,1,~obj.lqm.desired),3), 1, 1, 1, nRBFreq);

The signal power calculation divides the total transmit power equally across all RBs (transmitPower/nRBFreq), while the interference power is extracted from LQM (which is the scheduled power). I understand that this matches the platform’s default behavior where power is equally distributed

However, in my development work, I’m implementing differentiated power allocation where:

  • Different power levels are assigned to different frequency bands
  • Potentially different power per RB

My question is: Should the CLSM feedback SINR calculation (both S and I components):

  1. Continue using the equal power allocation assumption (transmitPower/nRBFreq) as in the current implementation for the whole baseband?
  2. Or should it use the actual scheduled power per RB as defined in my power allocation scheme for both signal power and interference power only for the scheduled RBs?

Could you please clarify what would be the correct approach when implementing non-uniform power allocation?

Thank you for your guidance.

Best regards,
Nick Jay

2

Dear Nick Jay,

as discussed in the post PMI Calculation, knowledge about the power allocation of the transmission is not available in the calculation of the CLSM feedback. The power allocation for transmission is set after the scheduling of the users, which can use the information from the CLSM feedback for the scheduling information.

Therefore, it is likely that you have no other choice than to assume uniform power allocation in the calculation of the CLSM feedback. If the scheme you are implementing allows a more accurate estimation of the transmit power that will be scheduled, then it is advisable to use your own implementation of the more accurate transmit power to be scheduled.

I hope this helps, best regards,
Agnes