Reporting of Rank Indicator (RI) Conformance Test

Open Script

This example demonstrates how to measure the Rank Indicator (RI) reporting performance using the LTE Toolbox™ under conformance test conditions as defined in TS36.101 Section 9.5.1.1 [ 1 ].

Introduction

This example highlights the use of the lteRISelect function which provides estimation of the RI. The performance of the RI estimation is also tested. The performance requirement defined in TS36.101 Section 9.5.1.1 [ 1 ] is as follows:

the ratio of the throughput obtained when transmitting based on UE reported RI and that obtained when transmitting with fixed rank 2 shall be >= 1 (Test 1)

the ratio of the throughput obtained when transmitting based on UE reported RI and that obtained when transmitting with fixed rank 1 shall be >= 1.05 (Test 2)

This example tests that these requirements are met.

Simulation Configuration

The example is executed for a simulation length of 10 frames at an SNR of 0.0dB. The fixed RI value is configured to be 2, therefore simulating TS36.101 Section 9.5.1.1 Test 1 [ 1 ]. A large number of NFrames should be used to produce meaningful results. The variable FixedRI controls which of the performance requirements described in the introduction are tested: FixedRI=2 corresponds to Test 1 and FixedRI=1 corresponds to Test 2. Note that a target SNR of 20.0dB rather than 0.0dB applies for Test 2.

NFrames = 10;
SNRdB = 0.0;
FixedRI = 2;

Propagation Channel Model Configuration

The structure, channel, contains the channel model configuration parameters. Note that TS36.101 Section 9.5.1.1 Test 3 [ 1 ] can also be implemented, by setting channel.MIMOCorrelation = 'High' (a target SNR of 20.0dB applies to Test 3, and the throughput ratio target is different).

channel.Seed = 1;                   % Channel seed
channel.NRxAnts = 2;                % 2 receive antennas
channel.DelayProfile = 'EPA';       % Delay profile
channel.DopplerFreq = 5.0;          % Doppler frequency
channel.MIMOCorrelation = 'Low';    % Multi-antenna correlation
channel.ModelType = 'GMEDS';        % Rayleigh fading model type
channel.NormalizeTxAnts = 'On';     % Normalize for transmit antennas
channel.NormalizePathGains = 'On';  % Normalize delay profile power
channel.InitPhase = 'Random';       % Random initial phases
channel.NTerms = 16;                % Oscillators used in fading model

Channel Estimator Configuration

The channel estimator is configured with a structure cec. The variable perfectChanEstimator controls channel estimator behavior. Valid values are true or false. When set to true a perfect channel estimator is used otherwise an imperfect estimate of the channel is used, based on the values of received pilot signals. In this example, we enable the perfect channel estimator.

% Configure channel estimator
cec.PilotAverage = 'UserDefined';   % Type of pilot symbol averaging
cec.FreqWindow = 9;                 % Frequency window size in REs
cec.TimeWindow = 9;                 % Time window size in REs
cec.InterpType = 'Cubic';           % 2D interpolation type
cec.InterpWindow = 'Centered';      % Interpolation window type
cec.InterpWinSize = 1;              % Interpolation window size

% Channel estimator behavior
perfectChanEstimator = true;

Set RI and CQI/PMI Delays

Initialize RI and CQI/PMI reporting delays in subframes; the RI delay is 9 subframes rather than 8 subframes in accordance with Note 5 of TS36.101 Table 9.5.1.1-1 [ 1 ]. Note that the feedback of the RI, CQI and PMI is assumed to be perfect, with the values being fed back in buffers rather than being fed back in an uplink transmissions.

riDelay = 9; % subframes
cqipmiDelay = 8; % subframes

Codebook Subset Restriction Bitmaps

In order to control the PMI selection which underlies the rank selection, codebook subset restriction bitmaps are used as described in TS36.213 Section 7.2 [ 2 ]. codebookSubsetRI1 configures the codebook subset restriction to the two precoders for rank 1, used for FixedRI=1. codebookSubsetRI2 configures the codebook subset restriction to the single precoder of rank 2, used for FixedRI=2. codebookSubsetUEReported is the union of these codebook subset restrictions and allows the rank to be selected dynamically in the case of UE reported RI.

codebookSubsetRI1 = '000011';
codebookSubsetRI2 = '010000';
codebookSubsetUEReported = '010011';

System Processing

The main processing is split into two phases, configured via the riConfig loop variable. These phases implement the two measurements required in the performance test defined in TS36.101 Section 9.5.1.1 [ 1 ]:

Fixed RI. The first phase (riConfig=1) performs PDSCH transmission and reception where the rank is set to a fixed value (either 1 or 2, selected by the variable FixedRI above). The final throughput is recorded.

UE reported RI. In the second phase (riConfig=2), the number of transmission layers is selected on the basis of UE reported RI, with the RI being updated every 5 subframes. The final throughput is recorded. The throughput ratio (measuredGamma) between the fixed RI phase and the UE reported phase is calculated and checked against the specified performance requirement.

Note that CQI and PMI are also updated every 5 subframes during both simulation phases.

The processing is performed on a subframe by subframe basis using the following steps:

Select RI. The current RI is read from the oldest value in the RI buffer riBuffer. The current RI is used to configure the number of transmission layers and codewords for the PDSCH.

Select CQI/PMI. The current CQI and PMI are read from the oldest values in the CQI and PMI buffers cqiBuffer and pmiBuffer. The current PMI is used to configure the precoding for the PDSCH. Note that for RI=2, there are two CQI values, one for each codeword; for RI=1 there is one CQI value for the single codeword.

Select MCS according to CQI. The Modulation and Coding Scheme (MCS) index corresponding to the CQI is selected for each codeword by means of a lookup table defined in TS36.101 Table A.4-1 [ 1 ] CSI RMC RC.2 FDD (MCS.2).

Determine Transport Block Size and modulation order. The MCS index for each codeword is passed to the lteMCS function which calculates the corresponding Transport Block Size (TBS) index and modulation order; the lteTBS function is then used to calculate the TBS for each codeword from the TBS index and the number of resource blocks allocated to the PDSCH.

Transmit and receive waveform. Transport block data is generated for one or two codewords as appropriate and passed to lteRMCDLTool to create a transmitted downlink waveform. This waveform is then passed through a fading channel and AWGN noise is added. The received signal is synchronized and OFDM demodulated and channel estimation is performed.

Measure PDSCH throughput. The PDSCH and DL-SCH are decoded and the CRC pass/fail for each codeword is recorded to determine the data throughput.

Update RI. If an RI update is scheduled in this subframe, use the channel estimate to update the RI with the lteRISelect function. The updated RI value is recorded in the RI buffer. The codebook subset restriction bitmap ensures that the reported RI will remain fixed for the fixed RI phase but will be dynamically selected for the UE reported RI phase. If an RI update is not scheduled in this subframe, the previous RI value is reused.

Update CQI/PMI. If a CQI/PMI update is scheduled in this subframe, use the channel estimate to update the CQI and PMI with the lteCQISelect and ltePMISelect functions. The updated CQI and PMI values are recorded in the CQI and PMI buffers. If a CQI/PMI update is not scheduled in this subframe, the previous CQI and PMI values are reused.

% Check that fixed RI valid is either 1 or 2
if (~any(FixedRI==[1 2]))
    error('Fixed RI value specified must be 1 or 2.');
end

% For each RI configuration (fixed RI and UE reported):
for riConfig = 1:2

    if (riConfig==1)
        riConfigStr = sprintf('fixed RI=%d',FixedRI);
    else
        riConfigStr = 'UE reported RI';
    end
    fprintf('\nSimulating with %s at %gdB SNR for %d Frame(s)\n', ...
        riConfigStr,SNRdB,NFrames);

    % Set up eNodeB settings for 2 codewords
    enb = struct('RC','R.3');
    enb.CellRefP = 2;
    enb.CFI = 3;
    enb.OCNGPDSCHEnable = 'On';
    enb.TotSubframes = 1;
    enb.PDSCH.RVSeq = 0;
    enb.PDSCH.CSIMode = 'PUCCH 1-1';
    enb.PDSCH.TxScheme = 'SpatialMux';
    enb.PDSCH.NLayers = 2;
    enb.PDSCH.Rho = -3.0;
    enb.PDSCH.CSI = 'On';
    ncw = 2;
    enb = lteRMCDL(enb,ncw);

    % Set channel model sampling rate
    ofdmInfo = lteOFDMInfo(enb);
    channel.SamplingRate = ofdmInfo.SamplingRate;

    % Configure appropriate codebook subset restriction and target
    % throughput ratio
    if (riConfig==1)
        if (FixedRI==1)
            enb.PDSCH.CodebookSubset = codebookSubsetRI1;
            targetGamma = 1.05;
        else
            enb.PDSCH.CodebookSubset = codebookSubsetRI2;
            targetGamma = 1.0;
        end
    else
        enb.PDSCH.CodebookSubset = codebookSubsetUEReported;
    end

    % Initialize variables
    totalCRC = [];      % CRC values, used for throughput calculation
    totalTBS = [];      % TBS values, used for throughput calculation
    RIReport = [];      % reported RI values
    riXaxis = [];       % corresponding subframe numbers
    PMIReport = [];     % reported PMI values
    CQIReport = [];     % reported CQI values
    SINRReport = [];    % corresponding SINR values
    cqipmiXaxis = [];   % corresponding subframe numbers
    offsets = 0;        % Initialize frame offset value
    rng('default');     % Default random number generator seed

    % Initialize RI/PMI/CQI buffers; note that the PDSCH throughput
    % resulting from these initial values is ignored, as the throughput
    % recording waits until the buffers have been filled with meaningful
    % reports
    riBuffer = ones(1,riDelay);
    pmiBuffer = ones(1,cqipmiDelay);
    cqiBuffer = ones(ncw,cqipmiDelay);

    % For each subframe:
    for subframeNo = 0:(NFrames*10-1)

        % Update subframe number
        enb.NSubframe = mod(subframeNo,10);

        % Update PMI, reading the oldest value from the PMI buffer
        cqipmiPtr = mod(subframeNo,cqipmiDelay);
        enb.PDSCH.PMISet = pmiBuffer(cqipmiPtr+1);

        % Update number of transmission layers based on RI, reading the
        % oldest value from the RI buffer
        riPtr = mod(subframeNo,riDelay);
        enb.PDSCH.NLayers = riBuffer(riPtr+1);

        % Update MCS according to CQI, reading the oldest value from the
        % CQI buffer
        CQI = cqiBuffer(:,cqipmiPtr+1);
        % Select MCS according to CQI using TS36.101 Table A.4-1 CSI RMC
        % RC.2 FDD (MCS.2), which defines the relationship between CQI
        % indices and MCS indices
        IMCSTable = [-1 0 0 2 4 6 8 11 13 15 18 20 22 24 26 27];
        IMCS = IMCSTable(CQI+1);

        % Determine TBS and modulation order
        [ITBS,modulation] = lteMCS(IMCS);
        if (mod(enb.NSubframe,5)==0)
            TBS = [0; 0];
        else
            TBS = double(lteTBS(size(enb.PDSCH.PRBSet,1),ITBS));
        end
        enb.PDSCH.TrBlkSizes(:,enb.NSubframe+1) = TBS;

        % Determine if an RI or PMI/CQI update is required in this
        % subframe, according to reporting periodicity N_pd = 5ms and
        % configuration indices cqi-pmi-ConfigurationIndex = 6 and
        % ri-ConfigurationInd = 1 from TS36.101 Table 9.5.1.1-1
        riPeriod = 5; % periodicity N_pd
        riOffset = 3; % offset deriving from configuration indices
        riUpdate = (mod(subframeNo,riPeriod)==riOffset);
        cqipmiPeriod = 5; % periodicity N_pd
        cqipmiOffset = 4; % offset deriving from cqi-pmi-ConfigurationIndex
        cqipmiUpdate = (mod(subframeNo,cqipmiPeriod)==cqipmiOffset);

        % Establish if this subframe actually needs executed for PDSCH
        % reception, RI/PMI/CQI estimation or initial timing offset
        % estimation:
        firstRxSubframe = max(cqipmiDelay+cqipmiOffset,riDelay+riOffset);
        if((any(TBS) && subframeNo>=firstRxSubframe) || ...
                (riUpdate || cqipmiUpdate) || subframeNo==0)

            % Configure data for transmission
            % (1 codeword on 1 layer, 2 codewords on 2 layers)
            data = {randi([0 1],TBS(1),1)};
            if (enb.PDSCH.NLayers==2)
                data = {data{1} randi([0 1],TBS(2),1)};
            end
            enb.PDSCH.Modulation = modulation(1:numel(data));

            % Create OFDM resource grid containing RMC transmission and
            % perform OFDM modulation
            txWaveform = lteRMCDLTool(enb,data);

            % The initialization time for channel modeling is set each
            % subframe to simulate a continuously varying channel
            channel.InitTime = subframeNo/1000;

            % Pass data through the fading channel model.
            % An additional 25 samples are added to the end of the
            % waveform. These are to cover the range of delays expected
            % from the channel modeling (a combination of implementation
            % delay and channel delay spread)
            rxWaveform = lteFadingChannel(channel, ...
                            [txWaveform ; zeros(25,size(txWaveform,2))]);

            % Calculate noise gain including compensation for downlink
            % power allocation
            SNR = 10^((SNRdB-enb.PDSCH.Rho)/20);

            % Normalize noise power to take account of sampling rate, which
            % is a function of the IFFT size used in OFDM modulation, and
            % the number of antennas
            N0 = 1/(sqrt(2.0*enb.CellRefP*double(ofdmInfo.Nfft))*SNR);

            % Create additive white Gaussian noise
            noise = N0*complex(randn(size(rxWaveform)), ...
                                randn(size(rxWaveform)));

            % Add AWGN to the received time domain waveform
            rxWaveform = rxWaveform + noise;

            % Perform synchronization
            % An offset within the range of delays expected from the
            % channel modeling (a combination of implementation delay and
            % channel delay spread) indicates success
            if (mod(subframeNo,10)==0)
                offset = lteDLFrameOffset(enb,rxWaveform);
                if (offset > 25)
                    offset = offsets(end);
                end
                offsets = [offsets offset]; %#ok<AGROW>
            end
            rxWaveform = rxWaveform(1+offset:end,:);

            % Perform OFDM demodulation on the received data to create the
            % received resource grid
            rxGrid = lteOFDMDemodulate(enb,rxWaveform);

            % Channel estimation
            if (perfectChanEstimator)
                chEstGrid = ...
                   lteDLPerfectChannelEstimate(enb,channel,offset);
                n = lteOFDMDemodulate(enb,noise(1+offset:end,:));
                noiseEst = var(reshape(n,numel(n),1));
            else
                [chEstGrid,noiseEst] = ...
                    lteDLChannelEstimate(enb,enb.PDSCH, ...
                        cec,rxGrid); %#ok<UNRCH>
            end

            % If this subframe requires PDSCH reception:
            if(any(TBS) && subframeNo>=firstRxSubframe)

                % Decode the PDSCH
                ind = ltePDSCHIndices(enb,enb.PDSCH,enb.PDSCH.PRBSet);
                pdschRx = lteExtractResources(ind,rxGrid) * ...
                            (10^(-enb.PDSCH.Rho/20));
                pdschChEst = lteExtractResources(ind,chEstGrid);
                [rxBits,rxSymbols] = ltePDSCHDecode(enb,enb.PDSCH, ...
                                        pdschRx,pdschChEst,noiseEst);

                % Decode the DL-SCH
                [decbits,crc] = lteDLSCHDecode(enb,enb.PDSCH, ...
                    enb.PDSCH.TrBlkSizes(1:numel(rxBits),enb.NSubframe+1),rxBits);

                % Record the CRC and TBS values for final throughput
                % calculation
                totalCRC = [totalCRC crc];                  %#ok<AGROW>
                totalTBS = [totalTBS TBS(1:numel(crc)).'];  %#ok<AGROW>

            end

            % Update RI:
            if (riUpdate)

                % Update RI
                thisRI = lteRISelect(enb,enb.PDSCH,chEstGrid,noiseEst);

                % Feed the value back to UE (in a buffer)
                riBuffer(riPtr+1) = thisRI;

                % Record values for plotting
                if (riConfig==2)
                    RIReport = [RIReport thisRI];           %#ok<AGROW>
                    riXaxis = [riXaxis subframeNo];         %#ok<AGROW>
                end

            end

            % Update CQI and PMI:
            if (cqipmiUpdate)

                % Update PMI conditioned on the most recently reported RI,
                % or the fixed RI if RI reporting is not configured
                if (isempty(RIReport))
                    enb.PDSCH.NLayers = FixedRI;
                else
                    enb.PDSCH.NLayers = RIReport(end);
                end
                thisPMI = ltePMISelect(enb,enb.PDSCH,chEstGrid,noiseEst);

                % Update CQI conditioned on the most recently reported PMI
                % and RI; this includes configuring the number of codewords
                % for CQI selection based on the RI
                enb.PDSCH.PMISet = thisPMI;
                enb.PDSCH.NCodewords = enb.PDSCH.NLayers;
                [thisCQI,thisSINR] = ...
                    lteCQISelect(enb,enb.PDSCH,chEstGrid,noiseEst);

                % Feed the values back to UE (in buffers)
                pmiBuffer(:,cqipmiPtr+1) = thisPMI;
                % For CSIMode='PUCCH 1-1', the CQI for the second codeword
                % is reported as a differential from the first codeword;
                % here we convert to absolute CQI indices for both
                % codewords.
                if (numel(thisCQI)==2)
                    thisCQI(2) = thisCQI(1) - thisCQI(2);
                end
                % Transmit with lowest CQI=1 when CQI selection reports
                % CQI=0 (out of range)
                thisCQI(thisCQI==0) = 1;

                cqiBuffer(:,cqipmiPtr+1) = thisCQI;

                % Record values for plotting
                if (riConfig==2)
                    PMIReport = [PMIReport thisPMI];        %#ok<AGROW>
                    CQIReport = [CQIReport thisCQI(1)];     %#ok<AGROW>
                    SINRReport = [SINRReport thisSINR(1)];  %#ok<AGROW>
                    cqipmiXaxis = [cqipmiXaxis subframeNo]; %#ok<AGROW>
                end

            end

        end

        % For subframes where RI was not updated, re-use the previous value
        % in the buffer
        if (~riUpdate)
            riBuffer(riPtr+1) = riBuffer(mod(riPtr-1,riDelay)+1);
        end

        % For subframes where PMI/CQI were not updated, re-use the previous
        % values in the buffers
        if (~cqipmiUpdate)
            pmiBuffer(:,cqipmiPtr+1) = ...
                pmiBuffer(:,mod(cqipmiPtr-1,cqipmiDelay)+1);
            cqiBuffer(:,cqipmiPtr+1) = ...
                cqiBuffer(:,mod(cqipmiPtr-1,cqipmiDelay)+1);
        end

    end

    % Display results for the current RI configuration
    fprintf('\nResults with %s:\n',riConfigStr);
    tputTotal = sum(totalTBS);
    if (riConfig==1)

        % Compute and display throughput
        tputFixedRI = sum(totalTBS.*(1-totalCRC));
        fprintf('Throughput: %d bits (%0.2f%%)\n', ...
            tputFixedRI,tputFixedRI/tputTotal*100);

    else

        % Compute and display throughput
        tputUEReported = sum(totalTBS.*(1-totalCRC));
        fprintf('Throughput: %d bits (%0.2f%%)\n', ...
            tputUEReported,tputUEReported/tputTotal*100);

        % Compute and display throughput ratio
        measuredGamma = tputUEReported/tputFixedRI;
        fprintf(['Throughput ratio (gamma): %0.3f' ...
            ' (requirement is >= %0.2f)'],measuredGamma,targetGamma);

    end

end

Simulating with fixed RI=2 at 0dB SNR for 10 Frame(s)

Results with fixed RI=2:
Throughput: 203184 bits (100.00%)

Simulating with UE reported RI at 0dB SNR for 10 Frame(s)

Results with UE reported RI:
Throughput: 235352 bits (100.00%)
Throughput ratio (gamma): 1.158 (requirement is >= 1.00)

Plot Results

Two figures are produced. The first figure has two subplots: the first subplot shows the reported RI for each subframe; the second subplot shows the reported PMI for each subframe. The second figure also has two subplots: the first subplot shows the estimated SINR for each subframe; the second subplot shows the reported CQI for each subframe. These plots illustrate how the RI, PMI and CQI vary over time due to the fading channel.

figure;
subplot(2,1,1);
plot(riXaxis,RIReport,'bo-');
xlabel('Subframe number');
ylabel('Selected RI');
title('Selected RI versus subframe number');
hold on;
subplot(2,1,2);
plot(cqipmiXaxis,PMIReport,'ks-');
xlabel('Subframe number');
ylabel('Selected PMI');
title('Selected PMI versus subframe number');

figure;
subplot(2,1,1);
plot(cqipmiXaxis,SINRReport,'rx-');
xlabel('Subframe number');
ylabel('Estimated SINR (dB)');
title('Estimated SINR (1st codeword) versus subframe number');
hold on;
subplot(2,1,2);
plot(cqipmiXaxis,CQIReport,'bo-');
xlabel('Subframe number');
ylabel('Selected CQI Index');
title('Selected CQI Index (1st codeword) versus subframe number');

Selected Bibliography

3GPP TS 36.101 "User Equipment (UE) radio transmission and reception"
3GPP TS 36.213 "Physical layer procedures"