3GPP-Like Channel Models at Sub-Terahertz: Improved Cluster Generation and Experimental Verification

The 3GPP channel model for new radios, described in 3GPP TR 38.901, is one of the most widely used channel models in the cellular wireless community. Originally established at below-6GHz radio frequencies and gradually expanded to cover new frequencies, its applicability to the new frequencies has always been an interest of the community. Our comparison of the radio channel responses from the measurements and the original 3GPP channel model, along with their variants adapted to sub-Terahertz frequency at 142 GHz in indoor and outdoor environments, sheds light on the possible improvement of the model. The original model defines cluster azimuth angles such that stronger clusters are always concentrated around a reference direction. Meanwhile, the measured azimuth power spectrum indicates that strong clusters randomly come from any direction. An alternative approach to iteratively generate cluster angles is therefore proposed, which provides improved agreement of the angular spread and eigenvalue statistics between measured and generated channels, at the expense of increased computational complexity in generating clusters. Moreover, the number of rays has no significant impact on the eigenvalue statistics. Setting the number of clusters fixed results in inaccurate statistics of the number of eigenmodes in generated sparse multipath channels.