nrDCIEncode versus physical layer subcomponents

I expect the MATLAB function nrDCIEncode to return the same output as when the chain of nrCRCEncode, nrPolarEncode, and nrRateMatchPolar functions are used for encoding DCI. In other words, rate_matched_dci_bits in the following code is expected to be equal to dcicw. But it is not! Why?

A = 64;
poly = '24C';
rnti = 100;
E = 864;
L = 24;
n_max = 9;
iil = true;
ibil = false;
K = A + L;

dci_bits = randi([0 1] , A , 1 , 'int8');
crc_attached_dci_bits = nrCRCEncode(dci_bits , poly , rnti);
polar_encoded_dci_bits = nrPolarEncode(crc_attached_dci_bits , E , n_max , iil);
rate_matched_dci_bits = nrRateMatchPolar(polar_encoded_dci_bits , K , E , ibil);

dcicw = nrDCIEncode(dci_bits , rnti , E); 

Answer 1

[Saffan]

Hi Alireza,

The reason for the difference comes from the implementation of "nrDCIEncode" function. The implementation in this function is different from the way you are encoding in the following ways:

• The “nrDCIEncode” function contains an additional step of prepending 24 ones to the input bits before calculating the CRC.
• It then removes the ones after CRC is calculated before proceeding to channel coding and rate matching.

Additionally, it also includes a step for converting the 24 CRC bits to a MSB first decimal representation using a binary-to-decimal conversion formula.