Graduation Year


Document Type




Degree Name

Doctor of Philosophy (Ph.D.)

Degree Granting Department

Electrical Engineering

Major Professor

Kwang-Cheng Chen, Ph.D.

Committee Member

Nasir Ghani, Ph.D.

Committee Member

Randy Larsen, Ph.D.

Committee Member

Ismail Uysal, Ph.D.

Committee Member

Srinivas Katkoori, Ph.D.


Neural Network, Artificial Intelligence, Channel State Information, Data Driven, MmWave, Path Loss


Channel prediction is a mathematical predicting of the natural propagation of the signal that helps the receiver to approximate the affected signal, which plays an important role in highly mobile or dynamic channels. The standard wireless communication channel modeling can be facilitated by either deterministic or stochastic channel methodologies. The deterministic approach is based on the electromagnetic theories and every single object in that environment has to be known in that propagation space and an example of this method is ray tracing. While the stochastic modeling method is based on measurements that involve statistical distributions of the channel parameters and an example of this approach is Floating-Intrcept (FI) model. In other words, channel modeling uses mathematical parameters to obtain the effect of the channel medium. These effects cause the transmitted signal to be either destructive or constructive during the propagation. Where the main focus of this dissertation is how Artificial Intelligence will be used in channel modeling. Fifth-generation -5G- with massive MIMO, higher data rate, handover, and channel modeling become more and more complex with the new wireless generations than the traditional stochastic or deterministic approaches. Nowadays, traditional wireless communication channel modeling is considered an old fashion especially with new technologies era such as things that applies to MmWave. In this sense, researchers and academia looking forward to more effective methods that have less complexity and more accuracy. Emerging machine learning technology supplies a new direction to process big measurement data and traffic data toward the wireless channel. Thus, new novel strategies of channel learning are proposed to generate a model free of the wireless channel modeling by tacking these difficulties.

With the availability of high computational devices and data, Artificial Intelligence (AI) emerges to revolutionize system design for new radio 5G. The subcategories of AI involve machine learning, deep learning techniques such as supervise leaning methods will be used to predict the channel state information (CSI) of a variate of environments base on a certain dataset. The fundamentals of wireless communication systems concentrate on channel modeling particularly for new frequency bands such as MmWave. Machine learning can facilitate rapid channel modeling for 5G wireless communication systems due to the availability of partially relevant channel measurement data and models. When irregularity of the wireless channels leads to a complex methodology to achieve accurate models, appropriate machine learning methodologies explore to reduce the complexity and increase the accuracy. This dissertation demonstrates an introduction to alternative procedures beyond traditional channel modeling to enhance CSI prediction based on data-driven with the usage of AI techniques, to alleviate the dilemma of channel complexity and time-consuming process that the measurements take. An example of applying AI towards wireless channel modeling is applying regression techniques with measurement data of a certain scenario to successfully assist the prediction of the path loss model of a different operating environment.

The irregularity of the wireless channel leads to a complex wayside to achieve accurate models where new technology is required to accomplish the precise results with new technologies. Machine learning algorithms involve channel modeling to reduce complexity and increase the accuracy that reduces the number of measurements. Furthermore, researchers explore machine learning methods that can link wireless channel modeling in different systems. Due to a large number of operations and extensive measurements, the researcher tends to perform machine learning to enhances the channel modeling prediction. The aim of using the machine learning algorithms is to develop alternative techniques to estimate the received signal that’s usually get affected by the channel. Moreover, extract and develop useful information from channel measurement data in the wireless communication system. Lacking using machine learning (ML) techniques for mobile wireless channel models is overcomes throughout this dissertation. By applying ML algorithms such as classification techniques will be used to investigate the wireless channel modeling and compare the result of each model by using the interpretation of performance measures such as accuracy, precision, and the number of misclassifications.

Artificial intelligence (AI), particularly machine learning (ML), is widely studied to enable a system to learn of intelligence, predict and make an assessment instead of the needs of humans. Switching the traditional channel modeling to machine learning channel modeling still in its early stage. One of the main issues in current communication is accurate of prediction the channel parameters, whereas using machine learning techniques could enhance the prediction and reduce the complexity. ML can be used to predict and estimate the wireless channel parameters and examine large and small-scale fading including parameters such as path loss, delay path loss exponent, frequency, Doppler spread and random variable that explains the large scale fading.

Usually, the supervised learning can be divided into two main subjects which are the regression and classification learning. The regression method is considered continuous values whereas the classification is a discrete value. Both of these two subjects are useful in estimating the channel parameters such as the path loss component and the large-scale random variable. The error can be minimized if optimization techniques are involved or by modifying some machine learning algorithms. One of the topics in this dissertation is to assist base stations to select the optimum signal based on the availability of the CSI data. This approach emerges as an attractive technique in the radio access network (RAN) and link selection to result in the strongest propagated link becomes the critical technology to facilitate RAN using mmWave. By taking advantage of existing operating data and apply appropriate artificial neural networks (ANN) algorithms to alleviate severe fading in the mmWave band. Additionally, we applied classification techniques using ANN with multilayer perception to predict the path loss of multiple transmitted links and base on a certain loss level, and thus execute effective relay selection, which also recommends the handover to an appropriate path. ANN with multilayer perception is compared with other ML algorithms to demonstrate effectiveness for relay selection in 5G-NR. Thus, machine learning (ML) is a new way that will change the design, standardization, and optimization of the communication systems. ML techniques such as supervise leaning and unsupervised methods will be used to estimate the wireless channel parameters by inferring CSI based on data-driven since the propagation signal of communication systems fundamentals is focusing on channel modeling particularly for a new technology era such as MmWave.

Millimeter-wave supplies alternative frequency bands of wide bandwidth to better realize pillar technologies of enhanced mobile broadband (eMBB) and ultra-reliable and low-latency communication (uRLLC) for 5G - new radio (5G-NR). With the usage MmWave frequency band, propagated signals become weaker due to fading through the channel and there must be a way to predict the strongest signal in a quick way to avoid the delay and to assist the coverage. With the usage of AI techniques and based on data-driven, predicting the strong signal can be achieved.

This dissertation focuses on predicting channel state information based on data-driven and elaborates on how to overcome some wireless issues in the new era 5G using Artificial Intelligence. Thus, based on our investigation in this dissertation, we can conclude that applying artificial intelligence towards wireless channel modeling is a promising technique and should be implemented in current and future wireless communication systems.