I use Vienna 5G LLS. I use NOMA scenario. there are two two method for channel estimation in this scenario. It has been observed that when using the two methods, the throughput of the users decreases and the throughput of the user becomes near from BS less than user in edge cell when using the method PilotAided. Otherwise, in the method of Approximate -Perfect. From your perspective, I want to clarify the work of the two methods and the reason for the difference in throughput.
For Approximate-Perfect method it is assumed that the channel is known and pilots are added to account for the overhead. The method PilotAided, on the other hand, uses the pilots to estimate the channel. Therefore, at low SNR (high channel estmation error) the throughput of the Approximate-Perfect scenario should be higher than the Pilot-Aided scenario and as the SNR increase (estimation error decreases) the thoughput gap between the two scenarios should decrease.