Impact of Limiting PAPR in 5G Uplink When Channel Estimation Is Not Accurate

In 5G Massive Multiple Input Multiple Output systems, the base station has to deal with a huge amount of high dimension channel matrices in order to estimate the channel response in a short time. That is because, ideally, the channel should be estimated from the evaluation of a different channel matrix at every subcarrier. However, it is possible to save matrices by assigning several subcarriers to the same channel matrix, assuming that they will behave similarly, provided that the subcarriers' aggregated bandwidth does not exceed the channel coherence bandwidth. On the other hand, reducing the peak-to-average power ratio (PAPR) is an important issue in 5G terminals. Clipping the signal is an easy way to deal with PAPR, but it also produces distortion. We present a study of the combined effects of these two distortions in order to determine to what extent they can be tolerated and to what extent the distortion produced by inaccurate channel estimation can mask the distortion generated by moderate clipping. The study is carried out by simulating different scenarios with varying coherence bandwidth, 5G numerology, the number of subcarriers included within each channel matrix, and limiting PAPR by applying several degrees of clipping.