Data-driven network design suggests that substantial technology advances of 5G and 6G networks will be enabled with enhanced automation, intelligence, and user-experience-focused capabilities. Network operators need to upgrade the standard network models by applying machine learning (ML) to address the complexities of next-generation network deployments. This article explores the role of ML and its interplay with wireless communications networks to develop the next-generation network architecture. A use case scenario for self-configuration of radio-access-network-based notification areas (RNAs) for effective resource management is analyzed to exemplify the proposed architecture where a paging load reduction of 64 percent is observed in the resulting RNA clusters. A conceptual framework for RNA configuration and management enabling a broader perspective toward an ML-driven hybrid self-organizing network is discussed to improve the signaling load to attain reduced latency and improved network capacity.

Non-orthogonal multiple access (NOMA) has been proposed as a promising multiple access (MA) technique in order to meet the requirements for fifth generation (5G) communications and to enhance the performance in internet of things (IoT) networks by enabling massive connectivity, high throughput, and low latency. This paper investigates the bit error rate (BER) performance of two-user uplink power-domain NOMA with a successive interference cancellation (SIC) receiver and taking into account channel estimation errors. The analysis considers two scenarios: perfect (ideal) channel estimation and a channel with estimation errors for various modulations schemes, BPSK, QPSK, and 16-QAM. The simulation results show that, as expected, increasing of the modulation level increases the SIC receiver BER. For example, at a signal-to-noise ratio (SNR) of 5 dB for perfect channel estimation and QPSK modulation, the user that is detected first has a BER of 0.005 compared to 0.14 for the user that is detected with the aid of the SIC receiver. Similarly, the BER of QPSK, assuming 0.25 channel estimation error of user 1, is equal to 0.06 at SNR = 15 dB compared to 0.017 for perfect estimation.

Providing high date rates that are independent of user location in the network is one of the Fifth Generation (5G) wireless network goals. This goal becomes even more challenging when the mobility of users is taken into account. The handovers that happens as the user crosses from cell to another can cause a sever degradation in the user's perceived data rate. The concept of Virtual Cell (VC) that is based on Coordinated Multipoint (CoMP) transmission is a promising solution for providing high data rates independent of user location in the network, and in particular for cell edge users. Multiple base-stations (BS) can coordinate with each other creating a Virtual Cell (VC). Users can roam within a Virtual Cell (VC) without the need to perform a handover. This diminishes the number of handovers a user will encounter, and also enhances the data rate for cell edge user by mitigating the Inter-Cell Interference (ICI) within a virtual cell. However, to enable the concept of Virtual Cells (VCs) rapid decisions need to be taken about when to enable / disable the VC mode and which Base Stations should be joining/leaving the VC as the user roams in the network. In this paper, the performance analysis of a novel algorithm based on a modification in the hidden layer of the Recurrent Neural Network (RNN), referred to as Gated Recurrent Units (GRUs). The RNN-GRU model is used for predicting the triggering conditions on enabling / disabling the VC mode. Sequences of the Received Signal Strength (RSS) values for different users in the network, were used for training the RNN-GRU model. After training, the RNN-GRU was used to predict the future RSS values, which is then used for making proactive decisions on enabling/disabling VC mode. Simulation results demonstrates that the proposed GRU-RNN model achieves an accuracy of 92% to predict the triggering conditions for enabling and disabling the CoMP mode as required based on the mobility of users.

Current procedures for anomaly detection in self-organizing mobile communication networks use network-centric approaches to identify dysfunctional serving nodes. In this paper, a user-centric approach and a novel methodology for anomaly detection is proposed, where the Quality of Experience (QoE) metric is used to evaluate the end-user experience. The system model demonstrates how dysfunctional serving eNodeBs are successfully detected by implementing a parametric QoE model using machine learning for prediction of user QoE in a network scenario created by the ns-3 network simulator. This approach can play a vital role in the future ultra-dense and green mobile communication networks that are expected to be both self-organizing and self-healing.

Diversity and Network Coding (DC-NC), which is the synergistic combination of Diversity Coding and Network Coding, was recently introduced to provide near instant link failure recovery and efficient transmission in a wide variety of network architectures. In this paper, DC-NC is applied to a multi-hop Wireless Sensor Network (WSN) and Diversity Coding (DC) is applied within each link in a multi-hop WSN to enable extremely high reliability with very rapid recovery from wireless link/node failures and the loss of data stream(s) at each link and also provide efficient transmission, which are very important metrics for WSNs. Furthermore, the reliability and rapid fault recovery can be extended to broadcasting applications, along with increased throughput.

Existing mobile networking systems lack the level of intelligence, scalability, and autonomous adaptability required to optimally enable next-generation networks like 5G and beyond, which are expected to be Self -Organizing Networks (SONs).It is anticipated that machine learning (ML) will be instrumental in designing future "x"G SON networks with their demanding Quality of Experience (QoE) requirements.This paper evaluates a methodology that uses supervised machine learning to predict the QoE level of the end user experiences and uses this information to detect anomalous behavior of dysfunctional network nodes (eNodeBs/base stations) in self-organizing mobile networks.An end-to-end network scenario is created using the network simulator ns-3, where end users interact with a remote host that is accessed over the Internet to run the most commonly used applications like file downloads and uploads and the resulting output is used as a dataset to implement ML algorithms for QoE prediction and eNodeB (eNB) anomaly detection.Three ML algorithms were implemented and compared to study their effectiveness and the scalability of the methodology.In the test network, an accuracy score greater than 99% is achieved using the ML algorithms.As suggested by the ns-3 simulation the use of ML for QoE prediction will help network operators understand end-user needs and identify network elements that are failing and need attention and recovery.

With the ever-increasing rise of a wide range of data-driven applications and services, as well as the synergies of gigabit wireless connectivity and pervasive broadband connectivity, there is a need for a paradigm shift in network methodologies to develop and deploy networks, such as 5G wireless. User-centric approaches to implementing self-organizing networks (SON) using machine learning (ML) have the potential to address the above challenges for 5G wireless communications networks and provide a seamlessly connected eco-system with superior user experience. This paper focuses on the potential performance improvements that can be achieved by integrating self-organizing networks and machine learning using user-centric approaches, with a focus on self-healing and self-optimizing SON functions.

Machine learning is expected to be a key enabler in 5G wireless self-organizing networks (SONs) that will be significantly more autonomous, smarter, adaptable and user-centric than current networks. This paper proposes a methodology, User Specific-Optimal Capacity Shortest Path (US-OCSP) routing, that uses machine learning to determine the resource-based optimum-capacity shortest path for a user between source and destination. The methodology takes into account two primary metrics, available capacity at network nodes (eNodeBs/gNodeBs) and distance, that are critical in determining the optimal path for an end-user. An ns-3 simulation determines the capacity, which is measured by the availability of resources [i.e., Physical Resource Blocks (PRBs)] at all possible serving network nodes between the source and destination, that is followed by implementation of Q-learning, a reinforcement type of machine learning algorithm that determines the shortest path avoiding congested network nodes so as to achieve the required throughput and/or bit rate. The ability to determine the optimal-capacity shortest path route will facilitate effective resource allocation that will optimize end-user satisfaction in a 5G SON network.

Mobile networking to achieve the ultra-low latency goal of 1 msec enables massive operation of autonomous vehicles and other intelligent mobile machines, and emerges as one of the most critical technologies beyond 5G mobile communications and state-of-the-art vehicular networks. Introducing fog computing and proactive network association, realizing virtual cell by integrating open-loop radio transmission and error control, and innovating anticipatory mobility management through machine learning, opens a new avenue toward ultra-low latency mobile networking.

This paper describes Slotted Aloha-NOMA (SAN), a novel medium access control (MAC) protocol, directed to Machine to Machine (M2M) communication applications in the 5G Internet of Things (IoT) networks. SAN is matched to the low-complexity implementation and sporadic traffic requirements of M2M applications. Substantial throughput gains are achieved by enhancing Slotted Aloha with non-orthogonal multiple access (NOMA) and a Successive Interference Cancellation (SIC) receiver that can simultaneously detect multiple transmitted signals using power domain multiplexing. The gateway SAN receiver adaptively learns the number of active devices using a form of multi-hypothesis testing and a novel procedure enables the transmitters to independently select distinct power levels. Simulation results show that the throughput of SAN exceeds that of conventional Slotted Aloha by 0.8 packet/sec and that of CSMA/CA by 0.2 packet/sec with a probability of transmission of 0.03, with a slightly increased average delay owing to the novel power level selection mechanism. Multiple-input-multipleoutput beamforming (2x2 MIMO BF) further increases the data throughput to 1.31 packet/sec compared with 0.36 packet/sec in conventional Slotted Aloha with 3 optimum power levels and reduces the average channel access delay of the SAN protocol from 0.13 sec to 0.09 sec.

In this paper the channel capacity for a dynamic random waypoint (RWP) mobility model of a Rayleigh fading channel is derived. A maximum ratio combining (MRC) diversity receiver and the effect of the number of branches, N, on the channel capacity is determined. As expected, by increasing the number of diversity branches, the resulting channel capacity is increased until the capacity is saturated. For example, the channel capacity for N = 3 is increased by 38.3% compared to no diversity ( N = 1 ) for the same value of the average received signal-to-noise ratio $\left( {\overline {{\text{SNR}}} } \right)$ but increasing N beyond 12 provides minimal gain. The channel capacity is compared with the classic Shannon capacity of the AWGN channel and with the well-known static model Rayleigh fading channel capacity. The channel capacity of the RWP Rayleigh channel is reduced by 10% compared to the AWGN Shannon capacity for a $\overline {{\text{SNR}}} $ of 20 dB. As expected, the AWGN channel capacity has a larger channel capacity as it is not affected as severely by fading as in the RWP mobility model. By contrast, the RWP model shows a slight improvement in channel capacity in comparison with the static model Rayleigh fading channel, since it will not be affected by severe fading for as long a time period as the static Rayleigh model. For example, the proposed model channel capacity increases to 6.11 bps/Hz whereas it is 5.87 bps/Hz for the static model Rayleigh fading channel capacity at the same $\overline {{\text{SNR}}} = 20{\text{dB}}$ with increasing of 4%.

The synergistic combination of Diversity and Network Coding (DC-NC) was previously introduced to provide very low end-to-end latency in recovering from a link failure and improve the throughput for a wide variety of network architectures. This paper is directed towards further improving DC-NC to be able to tolerate multiple, simultaneous link failures with less computational complexity. In this way, reliability will be maximized and the recovery time from multiple link or node failures is reduced in 5G fronthaul wireless networks. This is accomplished by modifying Triangular Network Coding (TNC) to create enhanced DC-NC (eDC-NC) that is applied to 5G wireless Fog computing-based Radio Access Networks (Fog-RAN). Our results show that using eDC-NC coding in Fog-RAN fronthaul network will provide ultra-reliability and enable near-instantaneous fault recovery while retaining the throughput improvement feature of DC-NC. In addition, the scalability of eDC-NC coding is demonstrated. Furthermore, it is shown that the redundancy percentage for complete protection is always less than 50% for the practical cases that were evaluated.

Fog networking has recently received considerable attention from a theoretical perspective, but in order for such networks to be practical several open areas need to be addressed. This paper determines the optimum number of nodes that should be upgraded to fog nodes with additional computing capabilities in order to maximize the average data rate and minimize the transmission delay. The optimization is performed for a given set of wireless channel conditions and a fixed total number of network nodes. It will be shown that, having more or less fog nodes than the optimum degrades the data rate. The numerical results indicate that the average data rate can increase nearly an order of magnitude for an optimized number of fog nodes in case of shadowing and fading. It is further shown that the optimum number of fog nodes does not increase in direct proportion to the increase in the total number of nodes. Furthermore, the optimum number of fog nodes decreases when channels have high path loss exponents. These findings suggest that the fog nodes must be selected among those that have the highest computation capability for densely deployed networks and high path loss exponent channels.

The Internet of things (IoT), which is the network of physical devices embedded with sensors, actuators, and connectivity, is being accelerated into the mainstream by the emergence of 5G wireless networking and the support of Machine-to-Machine (M2M) communications. Due to the simplicity of IoT devices and their sporadic traffic in such application, a simple medium access control (MAC) protocol is needed to connect M2M devices to the Internet through a hub (IoT gateway). This paper compares two MAC protocols namely: the newly introduced slotted Aloha-NOMA protocol and the well-known carrier sensing multiple access with collision avoidance (CSMA/CA) protocol. The comparison is based on two metrics, the throughput and the average delay. Simulation results show that the throughput of slotted Aloha-NOMA is higher than CSMA/CA at low probability of transmission at the cost of increased average delay caused by novel power level selection mechanism.

This paper proposes a new medium access control (MAC) protocol for Internet of Things (IoT) applications incorporating pure ALOHA with power domain non-orthogonal multiple access (NOMA) in which the number of transmitters are not known as a priori information and estimated with multi-hypothesis testing. The proposed protocol referred to as ALOHA-NOMA is not only scalable, energy efficient and matched to the low complexity requirements of IoT devices, but it also significantly increases the throughput. Specifically, throughput is increased to 1.27 with ALOHA-NOMA when 5 users can be separated via a SIC (Successive Interference Cancellation) receiver in comparison to the classical result of 0.18 in pure ALOHA. The results further show that there is a greater than linear increase in throughput as the number of active IoT devices increases.

Rapid recovery from link failures was previously demonstrated via the synergistic combination of Diversity and Network Coding (DC-NC) for a wide variety of network architectures. In this paper, the DC-NC methodology is further enhanced to achieve near-instant recovery from multiple, simultaneous wireless link failures by modifying Triangular Network Coding (TNC) to create enhanced DC-NC (eDC-NC) that is applied to 5G wireless Fog computing based Radio Access Networks (F-RANs). In addition, an explicit algorithm for the eDC-NC decoding process is provided. Our results demonstrate that applying eDC-NC coding to a F-RAN fronthaul network will provide ultra-reliability, enable near-instantaneous fault recovery, and enhance the throughput by at least 20% (for three broadcasted data streams).

Very fast link failure recovery and high throughput can be achieved via the synergistic combination of Diversity and Network Coding (DC-NC), an open-loop coding technique, in a wide variety of network architectures. In this paper, DC-NC is applied to enable recovering from a wireless link/node failure in the fronthaul portion of downlink Coordinated Multi Point (CoMP) 5G wireless network in a C-RAN environment. Our results demonstrate that by utilizing DC-NC coding in CoMP systems, resource consumption is reduced by about one-third and ultra-reliability with near-instantaneous fault recovery is achieved.

Expanding the cellular ecosystem to support an immense number of connected devices and creating a platform that accommodates a wide range of emerging services of different traffic types and Quality of Service (QoS) metrics are among the 5G’s headline features. One of the key 5G performance metrics is ultra-low latency to enable new delay-sensitive use cases. Some network architectural amendments are proposed to achieve the 5G ultra-low latency objective. With these paradigm shifts in system architecture, it is of cardinal importance to rethink the cell selection / user association process to achieve substantial improvement in system performance over conventional maximum signal-to- interference plus noise ratio (Max-SINR) and cell range expansion (CRE) algorithms employed in Long Term Evolution-Advanced (LTE-Advanced). In this paper, a novel Bayesian cell selection / user association algorithm, incorporating the access nodes capabilities and the user equipment (UE) traffic type, is proposed in order to maximize the probability of proper association and consequently enhance the system performance in terms of achieved latency. Simulation results show that Bayesian game approach attains the 5G low end-to-end latency target with a probability exceeding 80%.

Millimeter wave (mmWave) is a key technology to support high data rate demands for 5G applications. Highly directional transmissions are crucial at these frequencies to compensate for high isotropic pathloss. This reliance on directional beamforming, however, makes the cell discovery (cell search) challenging since both base station (gNB) and user equipment (UE) jointly perform a search over angular space to locate potential beams to initiate communication. In the cell discovery phase, sequential beam sweeping is performed through the angular coverage region in order to transmit synchronization signals. The sweeping pattern can either be a linear rotation or a hopping pattern that makes use of additional information. This paper compares recently proposed beam sweeping pattern prediction, based on the dynamic distribution of user traffic, using a form of recurrent neural networks (RNNs) called a Gated Recurrent Unit (GRU), and random starting point sweeping to measure the synchronization delay distribution. Results show that user spatial distribution and their approximate location (direction) can be accurately predicted based on Call Detail Records (CDRs) data using a GRU, which is then used to calculate the sweeping pattern in the angular domain during cell search. Moreover, the proposed beam sweeping pattern prediction enable the UE to initially assess the gNB in approximately 0.41 of a complete scanning cycle with probability 0.9 in a sparsely distributed UE scenario.

The integration of slotted Aloha with power domain non-orthogonal multiple access (NOMA), dubbed slotted Aloha-NOMA (SAN) can emerge as an appealing MAC protocol to be used for Internet-of-Things (IoT) applications over 5G networks. In this paper, SAN is discussed, and its performance is evaluated in detail. The simulation results demonstrate that the maximum normalized throughput can be increased from 0.37, which is the case for slotted Aloha, to 1 by means of SAN. Specifically, this full throughput efficiency can be obtained at all low, medium and high network traffics. Besides that, the average delay can be significantly reduced compared to the slotted Aloha.