Page 66 - Kaleidoscope Academic Conference Proceedings 2021

2021 ITU Kaleidoscope Academic Conference

range of APs × antennas = 256 to evaluate the performance. To realize CF-mMIMO in a real system, there are still many
The number of UEs is selected to be 16 as an example, issues to be considered. For example, as the number of
assuming that if the number of UEs is extremely small, the antenna sites and their antennas increases, the number of
results will be greatly affected by the location of the UEs, reference signals transmitted by the antennas of the antenna
while if there are too many, the interference between UEs sites increases, which increases the interference between the
will increase and the effect of the difference in AP reference signals and the overhead caused by those signals.
deployment on the results will be small. In addition, when a UE moves in an area where a large
number of antenna sites are distributed, reference signals
Table 1 – Simulation conditions from multiple antenna sites that the UE has to measure need
to be allocated so that they do not conflict with each other.
Parameter Value Another issue to be considered is how the transmit power
Carrier frequency 3.5 GHz control of the uplink should be performed.
Bandwidth 20 MHz 3.2 Key technology 2: User-centric RAN management
Transmission power 23 dBm
Simulation area 1km square The Beyond 5G system will provide various mobile
Number of antennas 256 (Number of APs x communications in a user-centric manner which realizes
Number of antennas per AP) continuous mobile communications with the user’s quality
Number of users 16 requirements anytime, anywhere. That is, networks are
multiplexed created and managed for each user in the Beyond 5G system.
Signal detection Zero Forcing From the viewpoint of wireless links, the CF-mMIMO in
Channel estimation Least Square Estimation Section 3.1 and the virtualized terminals in Section 3.3 are
the key technologies to realize a radio quality
homogenization for the wireless links. To realize a user-
centric RAN with the key technologies of wireless links and
the user-centric manner, a management scheme will be
designed to satisfy requirements of the key technologies.

Figure 4 – shows a difference of RAN structure between the
5G advanced of 2025 [11] and the Beyond 5G of the 2030s.
In the 5G advanced RAN of Figure 4 –(a), the RAN
virtualized technology, which is one of the enablers for the
advanced 5G RAN, will be applied to manage mobile
communication networks flexibly. A cell, which is the radio
coverage of a Radio Unit (RU), is fixed in the same way as
it is in the 5G system. RAN functions, that are the processing
of data and radio signals as a Central Unit (CU) and a
Distribution Unit (DU), are virtualized. The CUs and DUs
are placed to adapt mobile communication services. The
virtualized RAN technology can create multiple logical
networks per the services called RAN slices on the same
physical infrastructure such as transport links and general-
purpose computers. That is, each RAN slice has the RAN
Figure 3 – UE spectrum efficiency distributions [10] functions of vCU and vDU. Hereafter, the first letter “v”
before the acronyms means “virtualized” unless otherwise
Figure shows the simulation results of the distribution of the noted such as in Figure 4 – and Figure 5. The RAN
UE spectrum efficiency for each combination of the number Intelligent Controller (RIC) is defined by the O-RAN
of APs and the number of antennas per AP. The results show alliance [12], and it manages the creation of RAN slices and
that the UE spectrum efficiency is worse when all antennas the placement of virtualized RAN functions.
are placed in the same location (L=1, N=256). The UE
spectrum efficiency can be improved by distributing the Figure 4 –(b) shows our consideration of the user-centric
locations of the antennas. Among them, it is worth RAN in the Beyond 5G network. There are two major
mentioning that the UE spectrum efficiency results in the differences from the 5G advanced RAN: (1) cell-free
case of semi-distributed deployment, i.e., multiple antennas environment with CF-mMIMO, and (2) radio signal
in one location and multiple locations of those antennas are processing per users. The connected to one CPU and signals
better than the results of the UE spectrum efficiency when from all UEs were processed by the CPU. Recently, there has
all antennas are distributed (L=256, N=1). In other words, been active discussion on how to realize a mobile
this result indicates that there is a possibility of significantly communication system with CF-mMIMO based on the
reducing the number of antenna sites while maintaining the computing resources from initial study of CF-mMIMO [6]
effectiveness of distributed MIMO. mentioned that all APs were initial study of CF-mMIMO [6]

– 4 –