Recently, to cope with rapidly increasing data traffic, ITU-R established the concept of 5th generation mobile communication (5G) and defined requirements. Additionally, the 3GPP group has developed a new radio (NR) standard for 5G use, and the 5G era...
Recently, to cope with rapidly increasing data traffic, ITU-R established the concept of 5th generation mobile communication (5G) and defined requirements. Additionally, the 3GPP group has developed a new radio (NR) standard for 5G use, and the 5G era has opened in Korea.
Unlike current mobile telecommunication systems, the 5G mobile telecommunication system has high speed, ultra-low delay, super-connected characteristics, and it is planned to be used for IoT.
Also, the 5G will use the millimeter-wave band as an approach for ultra-high speed communication. The millimeter-wave band uses a wide frequency band. Thus, high data rates can be achieved.
However, high path loss, attenuation due to diffraction, etc. are significant challenges in using the millimeter-wave.
It is necessary to study the propagation characteristics of waves in detail to utilize the millimeter-wave in a communication system.
In particular, to represent the millimeter-wave characteristic in which the attenuation due to diffraction and transmission is very intense, it is necessary to analyze the influence of various obstacles on the millimeter-wave propagation characteristic.
Also, the 5G system considers the base station that can move to flexibly cope with the increasing data traffic and the disaster situation. Fast and accurate channel analysis and coverage prediction are essential for mobile base stations, such as drones.
The 5G and beyond system, meanwhile, plans to utilize the IoT platform actively. Estimating the location of devices is an essential aspect of increasing the practicality of IoT platforms.
In this thesis, I present techniques for resolving various issues that the beyond 5G system faces.
First, I proposed a method for improving the ranging performance for the localization using orthogonal frequency-division multiplexing-based communication system. The most difficult aspect of performing localization using the time-of-arrival information of a communication system is distinguishing indirect paths and noise from the direct path (DP) when the DP is blocked by obstacles. The combination of interference cancellation and an enhanced path detector is proposed to remove interference from nearby paths and detect low power DP. The proposed method is verified in 802.11ac environments, and it shows improved performance compared to conventional methods.
Next, I model roadside trees using a variety of techniques to analyze their effect on the accuracy of channel modeling using ray tracing simulation. The roadside trees were modeled as objects transmitting, reflecting, and diffracting electromagnetic waves. I noted that the results of ray tracing simulations that included roadside trees were more accurate than simulations without tree effects, based on the deviations from the experimentally measured results.
Finally, I propose a new algorithm for predicting the path loss exponent of outdoor millimeter-wave band channels through deep learning method. The proposed algorithm has the advantage of requiring less inference time compared to existing deterministic channel models while concretely considering the topographical characteristics.
I used three-dimensional ray tracing to generate the outdoor millimeter-wave band channel and path loss exponent. I trained a neural network with generated path loss exponent. To evaluate the performance of the proposed method, I analyzed the influence of the hyperparameters and environmental features, for example, building density and average distance from the transmitter.