Coexistence between 5G cellular and fixed satellite services in C-Band based on machine learning models
This thesis addresses limitation of existing Next-generation wireless mobile networks. The spectrum resource, especially below 6 GHz. such as 5th generation mobile networks, may reduce capacity bottlenecks by using radio frequency (RF) spectrum sharing. Spectrum sharing allows several wireless s...
Saved in:
| Main Author: | |
|---|---|
| Format: | Thesis |
| Language: | English |
| Published: |
2022
|
| Subjects: | |
| Online Access: | http://psasir.upm.edu.my/id/eprint/114908/1/114908.pdf http://psasir.upm.edu.my/id/eprint/114908/ http://ethesis.upm.edu.my/id/eprint/18207 |
| Tags: |
Add Tag
No Tags, Be the first to tag this record!
|
| Summary: | This thesis addresses limitation of existing Next-generation wireless mobile
networks. The spectrum resource, especially below 6 GHz. such as 5th
generation mobile networks, may reduce capacity bottlenecks by using radio
frequency (RF) spectrum sharing. Spectrum sharing allows several wireless
systems to coexist in a single spectrum band. The research on spectrum
coexistence difficulties between 5G base stations (BS) and fixed satellite
services (FSS) has recently increased. In Malaysia, the 5G uses frequencies
between 3.6 - 3.7 GHz, while FSS operates with 3.8 - 4.2 GHz. Although there
is a gap between both of them, adjacent channel interference might occur. For
this reason, the FSS downlink Earth Station (FSS-ES) interference should be
investigated from the 5G-BS to the FSS-ES and the 5G-ES to the 5G User
Equipment (UE). This is one of the main goals of this thesis, focusing on the
Malaysian scenario. Several proposals are drawn after realizing the interference
happened and affected the performance: The first part is the design of an
exclusion zone on how coexistence between 5G-BS and FSS-ES can be used
to avoid adjacent co-channel interference in 5G and B5G to FSS-ES. Co-channel
and adjacent channel interference are investigated at various stages in 5G-BS
and FSS-ES. For investigating interference in the same frequency band,
measurements have been carried out and data have been analyzed with 5G-BS
and FSS-ES. Then, 5G technologies addressed the optimal exclusion zone. In
order to analyze and improve state of the art, Machine Learning (ML) techniques
such as Radial Basis Function Neural Network (RBFNN) and General
Regression Neural Network (GRNN) have been used. The results indicated that
the proposed ML has its own set of characteristics that can be used to create a
new exclusion zone design that is more efficient. Furthermore, the adjacent
channel interference comprised the Interference-Noise Ratio (INR), where
interference occurred with INR levels below -12.2 dBm (-55dBc). It has been
shown that RBFNN has better accuracy, but lower MSE is obtained with GRNN.
The second part of the thesis focuses on the proposal of a filtering model
denoted Filter to Remove Broadband Interference 5G (FIREBRING) based on
the carrier-to-noise (C/N). It has to be designed jointly with the Guard Band (GB).
The results indicate that the proposed offered a complete analysis of the 5G
signal, considering the implications of out-of-band (OOB) emissions, potentially
LNB define saturation into the FSS receiver, and the repercussions of deploying
the 5G BS active antenna systems. With the LNB and down-converter in place,
it can be found that the signal interference between 1.450GHz and 1.550GHz, is
nearly 18dB. In the third part of the thesis, it is found that a lower look-up angle
for the FSS-ES is needed for future field trials with various 5G Active Antenna
Unit variants. The results suggest that 5G transmission operates at 3.620 GHz
to protect satellite services at 3.7 GHz. A further field trial was conducted to
evaluate further whether the distance and Guard Band (GB) can be reduced. It
is concluded that FSS-ES can coexist with 5G-BS as close as 85m apart, with
100 MHz GB and Bandpass Filter (BPF) rejection at least more than 45 dB. Also
includes a new filtering technique called 5G-Filter to Remove Interference in
Major Broadband (5G-FRIMB) to improve the signal. In the last part of the thesis,
an analytical model for 5G-BS and FSS-ES in C-Band based on ML for the
design of the exclusion zone is developed. In order to address these challenges,
this thesis examined whether it is possible to design a proper exclusion zone for
small cell 5G and FSS receivers based on the tropical region's characteristics.
Specific to the interference between 5G-BS and FSS-ES in the adjacent and cochannel
channel. Machine learning techniques have been used to model cochannel
interference. This PhD thesis shows that ML can help with some of the
modelling problems in RF, even in the presence of interference. |
|---|