Open Radio Access Network (O-RAN) is considered as a major step in the evolution of next-generation cellular networks given its support for open interfaces and utilization of artificial intelligence (AI) into the deployment, operation, and maintenance of RAN. However, due to the openness of the O-RAN architecture, such AI models are inherently vulnerable to various adversarial machine learning (ML) attacks, i.e., adversarial attacks which correspond to slight manipulation of the input to the ML model. In this work, we showcase the vulnerability of an example ML model used in O-RAN, and experimentally deploy it in the near-real time (near-RT) RAN intelligent controller (RIC). Our ML-based interference classifier xApp (extensible application in near-RT RIC) tries to classify the type of interference to mitigate the interference effect on the O-RAN system. We demonstrate the first-ever scenario of how such an xApp can be impacted through an adversarial attack by manipulating the data stored in a shared database inside the near-RT RIC. Through a rigorous performance analysis deployed on a laboratory O-RAN testbed, we evaluate the performance in terms of capacity and the prediction accuracy of the interference classifier xApp using both clean and perturbed data. We show that even small adversarial attacks can significantly decrease the accuracy of ML application in near-RT RIC, which can directly impact the performance of the entire O-RAN deployment.

Dynamic wireless resource allocation is one of the challenging problem in wireless networks to offer services with high data rate, high spectral efficiency, minimum energy and low latency when there is a high demand of limited wireless resources. To address this problem, we investigate an efficient wireless resource allocation scheme using fully connected convolutional neural network (CNN) with long time short memory (LSTM) algorithm for predicting demand of mobile virtual network operators (MVNOs) to sublease spectrum from wireless infrastructure providers (WIPs) to serve MVNOs' users. WIPs sublease wireless resources to MVNOs on the fly and MVNOs act like independent wireless service providers for their end users. While maximizing their payoffs and resource utilization, it is necessary for WIPs and MVNOs to allocate wireless resources adaptively to meet users' requirements. We present both formal mathematical analysis and numerical results to support our claims. Numerical results show that the WIP and MVNO make informed decisions based on prediction for wireless resources in order to maximum their payoffs when users change their requirements. Furthermore, performance comparison results show that the proposed approach outperforms the existing approaches in terms of payoffs, energy and delay.

Securing communication between any two wireless devices or users is challenging without compromising sensitive/personal data. To address this problem, we have developed an artificial intelligence (AI) algorithm to secure communication on virtualized wireless networks. To detect cyberattacks in a virtualized environment is challenging compared to traditional wireless networks setting. However, we successfully investigate an efficient cyberattack detection algorithm using an AI algorithm in a Bayesian learning model for detecting cyberattacks on the fly. We have studied the results of Random Forest and deep neural network (DNN) models to detect the cyberattacks on a virtualized wireless network, having considered the required transmission power as a threshold value to classify suspicious activities in our model. We present both formal mathematical analysis and numerical results to support our claims. The numerical results show our accuracy in detecting cyberattacks in the proposed Bayesian model is better than Random Forest and DNN models. We have also compared both models in terms of detection errors. The performance comparison results show our proposed approach outperforms existing approaches in detection accuracy, precision, and recall.

With the availability of different wireless networks in wireless virtualization, dynamic network selection in a given heterogeneous environment is challenging task when there is cyber security and data privacy requirements for wireless users. Selection of low cyber risk network can result in good service experience to the users. Network selection in virtualized wireless environment is determined by various factors such as Quality of Experience (QoE), data loss prevention, security and privacy. In this paper, we propose a learning model for dynamic network selection based on cyber-attack index (CI) value of networks. We have develop a recommendation system which recommends user to select the most secure network with least CI value. A mathematical model based on least squares and convex optimization is presented which predicts the CI of network with goal of maximizing the number of wireless users/subscribers. Numerical results show that the CI based recommendation system outperforms the traditional prediction based systems. Furthermore, we compare our approach with existing approaches and found that the proposed approach results in better performance in terms maximizing the number of wireless users/subscribers and better services to them.

In this paper, we investigate the machine learning approaches (sparse Bayesian linear regression (SBLR) and support vector machine (SVM)) for channel state information (CSI) prediction and dynamic radio frequency (RF) slicing for software defined virtual wireless networks in large-scale multi-input multi-output (MIMO) wireless networks. Specifically, a subset of the antennas of virtual wireless networks transmits pilot symbols for estimating the CSI and use the estimated CSI dataset to train and estimate the remaining channels and future CSI for virtual networks using machine learning algorithms. This helps not only to predict the CSI with least overhead and fulfills the service demands of users but also to reduce the power consumption and computation overhead in the network. Predicted CSI is leveraged for RF slicing for virtual wireless networks. Simulation results show that the proposed SBLR for predicting CSI results in lower BER and higher data rate for the wireless users. Furthermore, SBLR outperforms the other approaches when we have sparse CSI information and we need to generalize the prediction process.

In this article, we design, develop, and evaluate a wireless virtualization (WiVi) architecture for wireless networking for emerging Internet of Things (IoT) or 5G network applications by using a three-layer game theory model. WiVi is regarded as an emerging network paradigm to support different network service requirements (delay sensitive, bandwidth hungry, high data rate, etc.) for different IoT applications. In the proposed three-layer game model, interactions among three entities: 1) wireless resource providers (WRPs); 2) mobile virtual network operators (MVNOs); and 3) their subscribed wireless users (i.e., IoT devices), are formulated as strategies to optimize their respective utilities. The WiVi enables WRPs (also known as layer-1 leaders in the three-layer game) to sublease their wireless resources to MVNOs (also known as layer-2 leaders in the three-layer game) through RF slicing and adaptively setting their prices for subleasing. The MVNOs set the optimal competitive prices to attract more end-users/IoT devices (also known as followers in the three-layer game) to maximize their utilities. The end-users (IoT devices) maximize their data rates (i.e., utilities) by meeting the imposed quality of service (QoS) requirements and budget constraints. We investigate formal analysis of the uniqueness and existence of the equilibrium point of the three-layer game. Performance is evaluated using simulation results. Results show that the proposed three-layer game has unique and optimal equilibrium game. The numerical results show maximized utilities for WRPs, MVNOs, and IoT devices/users.

Malicious actions by cyber-adversaries are growing exponentially which makes it difficult to combat cyber-attacks for emerging networked cyber physical systems (CPS) and Internet of Things (IoT). Furthermore, wireless networks - major communication media for most emerging CPS and IoT applications - are highly vulnerable to cyber attacks because of their nature of open communications. In this paper, we evaluate the performance of the cyber deception system to combat cyber adversaries in virtualized wireless networking framework where software defined network (SDN) controller creates mobile virtual network operators (MVNOs) and continuously senses the network, observes the connections and creates deception MVNO to direct cyber adversaries. The deception MVNO can be used to learn about cyber adversaries in terms of their capabilities, intent and how much damage they can do in the system and so on. Thus, the cyber deception can help secure legitimate users from cyber adversaries. Performance of the proposed approach is evaluated with results obtained from Monte Carlo simulations.

In this paper, we study secrecy rate maximization through wireless virtualization in heterogeneous wireless networks where Mobile Virtual Network Operators (MVNOs) borrow frequency slices from Wireless Infrastructure Providers (WIPs) and serve the end users. For subleasing their frequency slices to MVNOs, WIPs use their service level agreements (SLA) with MVNOs to divide and reconfigure their RF bands. Then, we use two stage game for maximizing the payoff of MVNOs and secrecy rates (aka payoff) of end users. The performance of the proposed game is evaluated using numerical results where MVNOs offer lowers price to attract more users or more requests from users and the end users adapt their strategies to improve their secrecy rates.

The demand for digital media services is increasing as the number of wireless subscriptions is growing exponentially.In order to meet this growing need,mobile wireless networks have been advanced at a tremendous pace over recent days.However,the centralized architecture of existing mobile networks,with limited capacity and range of bandwidth of the radio access network and low bandwidth back-haul network,can not handle the exponentially increasing mobile traffic.Recently,we have seen the growth of new mechanisms of data caching and delivery methods through intermediate caching servers.In this paper,we present a survey on recent advances in mobile edge computing and content caching,including caching insertion and expulsion policies,the behavior of the caching system,and caching optimization based on wireless networks.Some of the important open challenges in mobile edge computing with content caching are identified and discussed.We have also compared edge,fog and cloud computing in terms of delay.Readers of this paper will get a thorough understanding of recent advances in mobile edge computing and content caching in mobile wireless networks.

In this paper, we study a machine learning enabled smart beam scheduling approach for wireless virtualization in large-scale multiple-input-multiple-output (MIMO) system. Large-scale MIMO is regarded as an emerging technology to enhance data rate of future wireless networks and the wireless virtualization is regarded as an efficient paradigm to enhance the radio frequency (RF) spectrum utilization by subleasing RF slices of wireless infrastructure providers to mobile virtual network operators (MVNOs). We leverage machine learning approach for scheduling the beams in large-scale MIMO where RF slices with the help of subsets of antennas are subleased for MVNOs. Performance of the proposed approach is evaluated using simulation results. The results show that the proposed approach outperforms the state of the art approach.

In this paper we present the experimental implementation of dynamic spectrum access (DSA) algorithm using universal software radio peripheral (USRP) and GNU Radio. The setup contains two primary users and two cognitive radios or secondary users. One primary user is fixed and the other is allowed to change its position randomly. Depending upon the position of the primary user the cognitive user will use the spectrum band where the detected energy is below certain predefined threshold level. The cognitive radio users are also programmed to operate independently without interfering with each other using energy detection algorithm for spectrum sensing. The modulation scheme is set to GMSK for secondary user performing data transmission. This experimental setup is used to analyze the quality of video transmission using DSA which provides the insight regarding the possibility of using free spectrum space to improve the performance of the system and its advantage over a non-DSA system. From the experiment it is shown that under congestion and interference DSA perform better than a non- DSA system.