Search

Radio maps can be used for source localization, link performance prediction, and wireless relay planning. This paper studies an air-to-ground radio map learning problem to predict the channel gain between a ground terminal and a low altitude unmanned aerial vehicle (UAV). The challenge is the insufficient measurement samples under the high dimensionality of the radio map, where both the transmitter and the receiver have full spatial degrees of freedom. Classical methods, such as k-nearest neighbor (KNN) and Kriging, may fail due to insufficient data. In this paper, a multi-degree channel model is proposed to fit low altitude air-to-ground propagation scenario. To address the high dimensionality issue, a hidden multi-class virtual obstacle model is developed, and a corresponding obstacle map is estimated from a small set of training data. Our numerical results confirm that the proposed model and the learning algorithm can significantly increase the prediction accuracy. When the constructed radio map applied to received signal strength (RSS) based source localization in a cellular communication scenario, a sub-20-meter accuracy is achieved, substantially better than the baseline model-based localization method.

In recent years, computation offloading has been considered as a promising technology to support computation- intensive vehicular applications. In this paper, we mainly focus on computation offloading in a vehicular cloud network (VCN), in which both vehicles and infrastructures with resource availability are defined as resource providers. However, due to the dynamically changing on-board resource distribution and the uncoordinated offloading strategies among vehicles, the computation offloading problem in a VCN is very challenging. In this paper we first model the problem as a multi-agent multi- armed bandit problem. We then propose a reshaped upper confidence bound (UCB) algorithm to estimate the on-board resource distribution with the reward estimation. We further utilize a novel multi-agent reinforcement learning algorithm to manage the computation offloading in a VCN. Simulation results demonstrate the performance gains by using the proposed algorithm.

This paper considers the problem of upgrading a legacy network into a Software Defined Network (SDN) over multiple stages and maximizing energy saving (ES) in the resulting upgraded network or hybrid SDN. In each stage, an operator needs to select and replace legacy switches with SDN switches and seek to switch off as many cables as possible over each link. This paper addresses the said problem where it considers (i) the available budget at each stage, (ii) maximum path delays, (iii) maximum link utilization, (iv) per-stage increase (decrease) in traffic size (upgrade cost), and (v) each non SDN switch must comply with the Open Shortest Path First (OSPF)-Equal Cost Multi-Path (ECMP) protocol. It outlines a Mixed Integer Program (MIP) and a heuristic algorithm called M-GMSU. The results show that (i) MIP and M-GMSU achieve ES of up to 71.93%, (ii) using a larger budget and/or number of stages increases ES, and (iii) the ES of M-GMSU is within 3.55% away from the optimal ES computed by MIP.

Unauthorized voice recording via smartphones can leak the talking content stealthily. This would be a serious security threat to those individuals, enterprises and the government who need to keep the conversation confidential. Furthermore, due to the size miniaturization of smartphones, it is hard to find the covert recording from malicious attendees. Existing solutions usually jam the recording with audible noise or electromagnetic emitting. However, the audible noise will seriously interfere with conversation and the effect of electromagnetic emitting will be limited by the distance. In this paper, we propose UltraArray, a pioneering silent ultrasonic anti-recording jammer, which can covertly block recording for a long distance. The principle of covert blocking is inspired by acoustic parametric array theory, which suggests that the audible frequency wave can be spread through the air silently while it is modulated to an inaudible ultrasonic frequency. The modulation used in this paper is double sideband (DSB) modulation. The microphone on the phone will record the audible frequency and filtering out the ultrasonic frequency. The jammer we developed uses an acoustics array to form a beam to spread the signal further. The evaluation shows that the device has a good jamming effect on more than 5 meters for most Android smartphones. It will also work well with more than 2.5 meters effective distance on iPhone XR, which has the active noise control (ANC) function. Those results achieve ten times the interference ability of existing solutions.

With the growing momentum around Open RAN (O-RAN) initiatives, performing dynamic Function Splitting (FS) in disaggregated and virtualized Radio Access Networks (vRANs), in an efficient way, is becoming highly important. An equally important efficiency demand is emerging from the energy consumption dimension of the RAN hardware and software. Supplying the RAN with Renewable Energy Sources (RESs) promises to boost the energy-efficiency. Yet, FS in such a dynamic setting, calls for intelligent mechanisms that can adapt to the varying conditions of the RES supply and the traffic load on the mobile network. In this paper, we propose a reinforcement learning (RL)-based dynamic function splitting (RLDFS) technique that decides on the function splits in an O-RAN to make the best use of RES supply and minimize operator costs. We also formulate an operational expenditure minimization problem. We evaluate the performance of the proposed approach on a real data set of solar irradiation and traffic rate variations. Our results show that the proposed RLDFS method makes effective use of RES and reduces the cost of an MNO. We also investigate the impact of the size of solar panels and batteries which may guide MNOs to decide on proper RES and battery sizing for their networks., Comment: Accepted Paper. 20XX IEEE. Personal use of this material is permitted. Permission from IEEE must be obtained for all other uses, in any current or future media, including reprinting/republishing this material for advertising or promotional purposes, creating new collective works, for resale or redistribution to servers or lists, or reuse of any copyrighted component of this work in other works

With the rapid development of the Internet of Things technology, smart speakers have become increasingly popular. However, smart speaker security is an ensuing threat. At present, smart speakers are activated by voice, and they monitor users’ voices 24 hours per day. Consequently, there may be problems with user privacy leakage. In this paper, we propose a non-intrusive digital forensic method for smart speakers. The main contribution of the paper is an effective method of combining network traffic analysis with the extraction of user intent and alarms about abnormal network traffic to support the investigation of security. We use Xiaomi smart speakers as an example in an experiment to verify our forensic method. The evaluation results show that our method works well for detecting security risks.

A Cyber-Physical System (CPS) refers to the interconnection of control (actuation), computational nodes and sensors, in order to manage physical processes. In recent years, the CPS design methodology has been adopted in several large-scale infrastructures such as smart power grids. Given the application criticality, sustaining the security of these systems is of utmost importance. One of the major security goals is to protect CPS against impersonation, where an adversary intends to manipulate the system state by sending erroneous data that appears to be reported by one of the system nodes, e.g. PMUs of a power grid. This paper proposes a novel hardware-assisted authentication scheme to counter such a threat, by exploiting imperfections that occur in the manufacturing process of integrated circuits. In essence, the proposed scheme associates a fingerprint for each system node so that the authenticity of the data source could be verified. In addition, the paper tackles the threat of message replay where the adversary re-transmits a legitimate message so that the system factors in outdated rather than fresh sensor measurements. This paper thwarts such a replay attack by leveraging the synchronized clocks across the CPS nodes, e.g., based on GPS; the idea is to employ a combination of time-stamp signatures and hardware fingerprints. Our proposed schemes can also detect and prevent data forgery, and Sybil attacks. The viability and performance of the proposed schemes are validated through analysis and prototype implementation.

The beyond fifth-generation (B5G) and future wireless networks face the challenges of spectral, energy and cost efficiency for large scale machine-type communications. Recently, emerging ambient backscatter communication (AmBC) technology provides a promising paradigm for the development of green Internet of Things (IoT) networks in beyond B5G era. In this paper, we consider a multi-nodes scenario where a backscatter network is symbiotic with primary network consisting of multiple ambient radio frequency (RF) sources, thereby allowing the system to use the appropriate RF to support high throughput and wide coverage for IoT devices. Unlike existing work on AmBC, which focuses on physical layer with relatively ideal model, i.e., classic three-nodes model composed of RF, backscatter device (BD) and IoT device, this paper studies the access control strategy, including coefficient design and device association of BDs and IoT devices, for multi-RF backscatter network from the perspective of maximizing device transmission rate. Under the guarantee of quality of service (QoS), we develop an access control strategy with aim of maximizing the weighted sum of primary and backscatter transmission rates, and design a distributed access control strategy called DCA-S, by using the difference of two convex functions approximation (DCA) and dual decomposition to transform the non-concave optimization problem into the solvable concave subproblems. Numerical results show that the proposed DCA-S can achieve significantly performance improvement of the system compared with benchmark schemes.

Terahertz frequency bands will likely be used for the next-generation wireless communication systems to provide data rates of hundreds of Gbps or even Tbps because of the wide swaths of unused and unexplored spectrum. This paper presents two outdoor wideband measurement campaigns in downtown Brooklyn (urban microcell environment) in the sub-THz band of 140 GHz with TX-RX separation distance up to 117.4 m: i) terrestrial urban microcell measurement campaign, and ii) rooftop surrogate satellite and backhaul measurement campaign. Outdoor omnidirectional and directional path loss models for both line-of-sight and non-line-of-sight scenarios, as well as foliage loss (signal attenuation through foliage), are provided at 140 GHz for urban microcell environments. These measurements and models provide an understanding of both the outdoor terrestrial (e.g., 6G cellular and backhaul) and non-terrestrial (e.g., satellite and unmanned aerial vehicle communications) wireless channels, and prove the feasibility of using THz frequency bands for outdoor fixed and mobile cellular communications. This paper can be used for future outdoor wireless system design at frequencies above 100 GHz.

In this paper we explore solutions to robust group communication in disadvantaged wireless networks that exhibit low bandwidth, high packet losses, and frequent or even permanent network partitions. More specifically, we propose a group communication protocol based on Named Data Networking (NDN). By design, NDN’s in-network caching and stateful forwarding plane can help improve data delivery robustness in disadvantaged networks, but when content is generated with dynamic rates, an efficient, low-overhead synchronization protocol is needed to inform group members to fetch newly generated content promptly. In this paper, we describe the Prefetch Loss-Insensitive Sync (PLI-Sync) protocol for group communication in highly disadvantaged networks. PLI-Sync combines optimistic content pre-fetching with new content notification via sync, and addresses the challenges of how to distinguish wireless packet losses from mobility-induced disconnections, and between data availability and retrievability. Our evaluations show that leveraging the interplay between optimistic pre-fetching and a low rate sync protocol can significantly reduce communication overhead compared to relying on sync protocols alone, while maintaining low data fetching latency and robust delivery in a variety of wireless conditions and traffic load settings.

This paper describes Mirror Pulse Modulation (MPM), which is an image sensor-based optical wireless communication protocol using line scan sampling. MPM is designed to be featured on current and future smartphones, and has the advantages of a wide communication range and the ability to transmit small pieces of information at high speed. MPM has been published as IEEE standard 802.15.7-2018, but this is the first paper to describe it in detail. We evaluate here the performance of MPM with parameter settings individually optimized for several scenarios. Our results show a 4.4-fold increase in communication range and a 2.3 fold increase in dimming rate compared to International Electrotechnical Commission (IEC) standard 62943:2017, which is adopted in several commercial services.

This paper aims to propose a three-dimensional (3D) point process that can be employed to generally deploy unmanned aerial vehicles (UAVs) in a large-scale cellular network and tractably analyze the fundamental network-wide performances of the network. This 3D point process is devised based on a 2D marked Poisson point process in which each point and its random mark uniquely correspond to the projection and the altitude of each point in the 3D point process, respectively. We elaborate on some important statistical properties of the proposed 3D point process and use them to tractably analyze the coverage performances of a UAV-enabled cellular network wherein all the UAVs equipped with multiple antennas are served as aerial base stations. The downlink coverage of the UAV-enabled cellular network is found and its closed-form results for some special cases are explicitly derived as well. Furthermore, the fundamental limits achieved by cell-free massive antenna array are characterized when coordinating all the UAVs to jointly perform non-coherent downlink transmission. These findings are validated by numerical simulation., 6 pages, 4 figures, conference paper. arXiv admin note: substantial text overlap with arXiv:2007.09866

This paper proposes solutions to channel prediction with atmospheric duct based on feedforward neural network (FNN) modeling. Specifically, FNN-based model is to produce accurate prediction by directly learning from large database rather than depending on any assumption. The prediction accuracy of the model applied to Sub-6 GHz and 28 GHz bands attain 88.72% and 94.77%, respectively. Besides, the paper also validates the difference of prediction performance of networks by comparisons of artificial neural network (ANN) and FNN-based networks. The results show that when the bands get higher, FNN-based framework would enhance the prediction accuracy while ANN-based framework would bring it down. It is demonstrated that the proposed FNN-based network obtain accuracy gain over 30% than ANN-based framework.

The fifth-generation (5G) mobile telecom network has been garnering interest in both academia and industry, with better flexibility and higher performance compared to previous generations. Along with functionality improvements, new attack vectors also made way. Network operators and regulatory organizations wish to have a more precise idea about the security posture of 5G environments. Meanwhile, various security metrics for IT environments have been around and attracted the community’s attention. However, 5G-specific factors are less taken into consideration.This paper considers such 5G-specific factors to identify potential gaps if existing security metrics are to be applied to the 5G environments. In light of the layered nature and multi-ownership, the paper proposes a new approach to the modular computation of security metrics based on cross-layer projection as a means of information sharing between layers. Finally, the proposed approach is evaluated through simulation.

Air-Ground Integrated Power Internet of Things (AGI-PIoT) is a key enabler to meet the stringent communication and computing requirements of PIoT devices. In AGI-PIoT, the computation-intensive and delay-sensitive tasks can be either offloaded to edge servers through unmanned aerial vehicles (UAVs) or offloaded to cloud servers through ground base stations (GBSs), while the computational resources of edge servers and cloud servers should be jointly allocated. However, the joint optimization of task offloading and resource allocation faces several challenges such as incomplete information, dimensionality curse, and coupling between long-term constraints of queuing delay and short-term decision making. In this paper, we propose a learning-based QUeue-AwaRe Task offloading and rEsouRce allocation algorithm (QUARTER). Specifically, by exploiting Lyapunov optimization, the joint optimization problem is decomposed into task offloading and server-side resource allocation. For the first subproblem, we propose a Queue-aware Actor-Critic-based task offloading algorithm named QAC to cope with dimensionality curse. A low-complexity heuristic algorithm is developed to solve the second subproblem. Compared with existing task offloading and resource allocation algorithms, simulation results demonstrate that QUARTER has superior performances in throughput, queuing delay, and convergence.

This paper aims to investigate the joint optimization of multidimensional constellation design (MCD) and multiuser detection (MUD) for grant-free non-orthogonal multiple access (GF-NOMA). We first formulate the joint optimization problem and derive its explicit expression using variational inference. Due to the intractability of the joint optimization problem, we then resort to deep learning (DL) and approximate the optimal solution in an end-to-end manner. Specifically, we develop a novel variational autoencoder based network, such that the distribution of the multidimensional constellations can be accessed and optimized. We also design a multi-task learning architecture on the decoder side to deal with the complex coupling among signal streams, by taking the MUD process as multiple distinctive yet related tasks. The derivation of the loss function for network training is presented, and simulation results are provided to validate the superior performance of the proposed method over conventional approaches.

Blockchain is a distributed ledger technology based on the underlying peer-to-peer network. In this paper, we focus on improving the chances of a transaction being packaged into a valid block so as to be recorded on the main chain. Blockchain nodes typically broadcast transactions they receive to the whole network. Hence, for recording transactions on the blockchain more efficiently, it becomes essential to determine influential nodes to publish transactions, where influential nodes are more actively involved in mining, recording, or broadcasting transactions in the network. To that aim, we propose an Influential Node Determination Framework (INDF) using a series of significant factors, such as hash rate, latency, active time, and degree of a node. Specifically, INDF consists of two parallel schemes: the first scheme figures out influential pools according to their hash rates where a truth-telling mechanism design is employed to encourage the pool nodes to report their true hash rate values; the second one determines influential individual nodes based on an improved L-H index algorithm. Remarkably, the proposed truth-telling mechanism is proved to be incentive-compatible. Our improved L-H index algorithm is comparatively studied to reflect the impacts of different node parameters on the node’s ranking. Extensive experiments are conducted to demonstrate the effectiveness of our proposed framework.

The rapid development of the mobile Internet makes it necessary to adopt efficient cryptographic primitives for the portable devices with limited computing resources. Online/offline identity-based signatures are suitable because of short response time of signature generation and being free from the cumbersome operations caused by public key infrastructures. In this paper, we propose the first identity-based chameleon hash which can be proved secure without the random oracle and show how to use it to translate any identity-based signature to an online/offline one.

Detailed systems engineering is required to integrate multiple self-interference cancellation (SIC) techniques for inband full-duplex (IBFD) wireless designs enabling novel applications. These SIC methods span the propagation, analog and digital domains of a typical transceiver, and must be individually analyzed before combining them to create an IBFD system. This paper discusses the various techniques within these domains as well as presents a survey of measured IBFD systems that have successfully demonstrated SIC across multiple domains. Additionally, several key research areas are outlined, and highlight the need for a resource that details both the fundamentals and state-of-the-art performance of IBFD techniques, systems and applications in order to accelerate the adoption of this technology within networking standards.

Energy efficiency (EE) is a crucial performance metric in wireless-powered backscatter communication networks (WP-BackComNets) for achieving a good tradeoff between data rates and the overall energy consumption, which however has not been sufficiently exploited by the existing works. In this paper, an EE-based maximization resource allocation (RA) problem is studied in a downlink orthogonal frequency division multiple access-based WP-BackComNet, where the circuit power consumption of the backscatter device, the minimum energy harvesting (EH) constraint, and the maximum transmit power constraint of the power station are considered. To deal with the non-convex problem, we firstly transform it into an equivalently subtractive form via Dinkelbach's method. Then, we apply a variable substitution approach to transform the non-convex problem into a convex one, where the closed-form solutions of the reflection coefficient, the transmit power, and the EH time are deduced by using Lagrange dual method. Simulation results demonstrate that the proposed algorithm can achieve better EE performance than other benchmark algorithms.

With the development of software-defined radio (SDR) and artificial intelligence (AI), smart jammers are becoming more and more powerful, and have cognitive and intelligent capacities such as spectrum sensing, learning and reconfigurability. To cope with the AI-enabled smart jammers, this paper investigates the naive weakness of smart jamming strategies and proposes a deceptively adversarial attack (DAA) based proactively intelligent anti-jamming algorithm to ensure reliable communication for wireless communication networks (WCNs). In contrast to the existing intelligent schemes, the proposed DAA-based anti-jamming algorithm focus on minimizing total reward of the smart jammer and destroy their spectrum sensing and learning abilities by delivering a carefully designed deceptively adversarial signal. We implement the DAA-based anti-jamming algorithm in both white-box and black-box settings according to the degree of information acquisition from the smart jammers, respectively. Simulation results show that system performance in terms of anti-jamming reward, computational complexity and anti-jamming efficiency can be significantly improved compared with the existing intelligent schemes.

In recent years, we have witnessed the viability of applying cloud computing concepts to the domain of vehicular networks. A basic component of this infrastructure derived from the merge of cloud computing and vehicular networks is a Vehicular Micro Cloud (VMC), also known as vehicular cloudlets. A VMC is a cluster of connected vehicles that share computational resources. Despite being the focus of many studies in recent years, we still do not have a clear understanding of the characteristics of VMCs in large-scale urban scenarios. In this paper, we investigate some fundamental characteristics of stationary and mobile VMCs obtained from a realistic vehicular mobility trace. We characterize the dwell time and the inter-arrival time in stationary VMCs. Also, using statistical modeling, we identify theoretical distributions that best fit these metrics. For mobile VMCs, we reveal how they occur throughout the city along the day, discussing evolution and lifetime aspects.

Cloud offloading enables smaller edge devices which contain fewer computing resources to be used for computer vision. This contributes to mobile applications of computer vision such as robotics and augmented reality. However, cloud offloading is impacted by bandwidth fluctuations in wireless networks. When the available bandwidth is restricted, it is difficult to offload workloads (i.e. frame) to cloud instances. Instead of offloading workloads to the cloud, existing approaches send a result of edge prediction or use a past cloud prediction to cover non-offloaded frames, which result in low accuracy. In this paper, we propose an Edge Cloud Ensemble method for object detection to improve accuracy in low bandwidth environments. In our method, the edge sends an inaccurate prediction and a motion vector of a detected object’s region to the cloud while maintaining low transmission overhead. The cloud corrects the inaccurate prediction by using the motion vectors to shift a past, accurate cloud prediction. The results of our experiments demonstrate that our approach can improve accuracy in low bandwidth environments compared with existing methods, especially in moving cameras.

In this paper, we propose an online learning approach, utilizing the framework of Online Convex Optimization (OCO) to tackle the problem of learning the optimal energy supply plan, in terms of total incurred cost, from the perspective of an electricity retailer that also owns power generators and renewable energy sources. The retailer does not have prior knowledge about key dynamic processes that affect the problem, such as future demand, renewable energy supply, and wholesale market prices. The retailer sequentially learns how to adjust the power generation plan so as to minimize the power generation cost plus the cost of buying additional power from the wholesale market, in case the planned generated amount is not enough to cover the demand. We use Online Mirror Descent (OMD) and Online Gradient Descent (OGD), and we verify that they both achieve sublinear static and dynamic regret, which compare the cumulative cost of each algorithm against that of the optimal offline static and dynamic solution respectively. In particular, dynamic regret appears to be well aligned with the considered setting since it captures the fact that a power generator can dynamically change its power output between consecutive time slots based on only small adjustments and not in an arbitrary fashion, due to ramp constraints. Our model can capture different settings of electricity markets. Simulations with real data verify that OMD precisely learns the optimal dynamic supply policy for small power adjustments between consecutive time slots.

In recent years, unmanned aerial vehicles (UAV) have been widely adopted to support complex target tracking tasks for military and civilian applications, especially in open and unknown environments. In practical cases, the moving trajectory of the target cannot be known to the UAVs in advance, which brings great challenges to UAVs to realize real-time and effective tracking. In addition, the limited tracking ability of a single UAV can hardly meet the requirements of a high tracking success rate. To deal with these problems above, this paper establishes a multi-UAV cooperative target tracking system. Besides, a deep reinforcement learning (DRL) based algorithm is designed to enable UAVs to make flight action decisions intelligently to track the moving air target, according to the past and current position information of the target only. To further increase the detection coverage of the UAV network when tracking, spatial information entropy is introduced to the reward designing in this algorithm. Simulation results validate that the proposed algorithm yields impressive target tracking performances, and significantly outperforms several common DRL baselines in terms of the tracking success rate. The convergence of the algorithm is also verified by the simulations.

In this paper, we focus on a low-powered short frame OFDM communication system where sensor network’s devices send only one OFDM symbol per frame. We study and model the design of this low-powered short frame OFDM communication system in terms of error correction. In order to respond favourably to the requirements of the low-powered sensor network, we analyse the performance of several Error Correcting Codes (ECCs); such as Hamming, Reed-Solomon (RS) and Convolutional Code (CC); based on different parameters (coding rates, complexity of implementation and correction capacity). We show that our design allows the sensor network communication model to operate close to theoretical limits. Moreover, we discuss the trade-off to make by choosing the best ECC (between Hamming, RS or CC) in order to maintain a low Bit Error Rate (BER) and to meet the requirements of the low-powered sensor network.

In this paper, we study the trajectory control and sub-channel assignment for unmanned aerial vehicle (UAV) based wireless networks with wireless backhauls. This design aims to maximize the min rate achieved by ground users (GUs) subject to their data transmission demands. To tackle the underlying mixed integer non-linear optimization problem, we use the alternating optimization approach where we iteratively optimize the sub-channel assignment and UAV trajectory control until convergence. Toward this end, we propose a heuristic algorithm to obtain a feasible solution for the sub-channel assignment sub-problem. In addition, the successive convex approximation (SCA) method is used to convexify and solve the non-convex UAV trajectory control sub-problem. Via extensive numerical studies, we illustrate the effectiveness of our proposed design for different network settings.

The travel delay caused by incidents severely reduces the efficiency of traffic. This symptom has been relieved with the development of advanced communication technologies. However, the emergency traffic information (ETI) is meaningless for vehicles which do not traverse the incident segment. Unlike most existing studies concentrating on the network performance during the dissemination of ETI to all vehicles, this paper proposes a joint traffic-communication optimization strategy (JTCS) to reduce the extra cost caused by unnecessary communication, which minimizes the total communication and traffic cost by transmitting the ETI to the worthy vehicles who need the ETI. Specifically, we capture the optimal targeted ETI transmission strategy with combination of radio resource allocation strategy (RRAS) and traffic-influencing transmission strategy (TTS), which can be converted into a bi-level optimization problem. The lower-level problem minimizes the total cost by Lagrangian method and obtains the optimal RRAS when the TTS is given. The upper-level problem develops the optimal TTS based on the optimal RRAS obtained in the lower-level problem. Simulation results using SUMO and MATLAB indicate that JTCS can achieve minimal total cost of ETI transmission and vehicle rerouting by comparing with existing approaches.

Traffic Matrix (TM) clearly describes the volume and the distribution of traffic flows inside a network. TM plays an important role in many network management fields, such as traffic accounting, short-time traffic scheduling or re-routing, network design, anomaly detection, etc. Hence, an accurate TM prediction strategy is essential to handle those tasks effectively. Fortunately, Artificial Intelligence (AI) has been developing very strongly, thanks to computer technology developments such as GPU and TPU. That offers an opportunity to apply AI to TM prediction methods. However, applying Machine Learning techniques in traditional networks encounters some issues due to the distributed control and restricted local view of network nodes. For this purpose, a centralized control architecture, e.g., Software-defined Network (SDN), is a promising candidate. In this paper, we apply Long Short-Term Memory (LSTM) and its two variants, Bidirectional LSTM (BiLSTM) and Gate Recurrent Unit (GRU), for TM prediction mechanisms of an SDN architecture network. The prediction models have been evaluated using two datasets: the popular GEANT backbone network traffic data and our dataset generated through a testbed. The experimental results show that our approach yielded promising traffic prediction accuracy.

Driven by the new era of big data and artificial intelligence (AI), as well as the increasing demands for the privacy protection, how to deployment the AI on wireless networks is drawing increasing attention. In this paper, we investigate the distributed learning mechanism of hosting AI over intelligent reflecting surface (IRS)-assisted wireless networks, where IRS is utilized to enhance communication in a cost-effective and energy-efficient manner. Firstly, a distributed learning framework is formulated based on the alternating direction method of multipliers (ADMM) to achieve the parallel processing of the objective function. Specifically, in the proposed architecture each user updates the learning model with its own data and uploads it to the global model through wireless networks. Thence, a joint passive phase shift of IRS and user scheduling scheme based on a metric of efficiency-efficacy weighted sum (EEWS) is formulated to explore both the learning efficiency and efficacy. In addition, aiming at improving the one-round learning efficiency, a grouping-based suboptimal solution about IRS’s phase is adopted to realize the max-min fair transmission. Simulation results demonstrate the relationship among the number of users involved, the scale of IRS and the learning performance.

Social Internet of Things (SIoT) has become an emerging network paradigm, where IoT devices with Artificial Intelligence (AI) and social relations can automatically establish a collaborative group to identify events locally. On the other hand, mobile users can act as ubiquitous and versatile sensors to improve the accuracy of SIoT event detection. In this paper, we explore the SIoT Collaboration with Crowdsourcing (SCC) problem to jointly select SIoT devices and hire users to monitor events and locations with accuracy requirements, while minimizing the total SIoT communication and computation costs and the user hiring cost. We prove that SCC is NP-hard and cannot be approximated by any factor unless P = NP. Then, we propose a new algorithm, Accuracy- and Social-aware SIoT and User Selection (ASSUS), with the idea of Collaborative Tree (CT) and Accuracy Profit (AP), where CT exploits users’ social relations to properly choose intermediate SIoTs. Simulation results manifest that ASSUS can effectively reduce more than 50% of the total cost compared with state-of-the-art algorithms.

With the rapid growth of Internet of Things (IoTs), Internet-connected devices and home appliances gain popularity on the consumer electronic market. New home IoT products with built-in network connections and intelligent functionalities are quickly rolled out to the market. As predicted by Gartner, there will be more than 500 IoT devices deployed in a typical household by 2022. The easy device integration and advanced automation logic also brings new challenges with regard to security and privacy. IoT devices have been reported as unreliable because of the constraints in costs and resources. Anomalies of IoT devices include malfunctions of the physical part or the cyber part of an IoT device, as well as abnormal behaviors due to malicious attacks. Abnormal IoT devices could cause severe consequences, because they reside in the home environment and have critical functions that can change the physical world, such as door (smart lock) opening, smart oven burning (which could cause fire), or smart water valve opening (which could cause flooding). In this paper, we study the important issue of localizing the root cause of anomalies in a smart environment (e.g., smart homes and smart offices). We propose to use decision trees for efficient and effective anomaly root cause localization. We construct decision trees from automation rules that control the operations of smart IoT devices in a smart environment. Our performance evaluation on data collected from real smart homes demonstrate the effectiveness of our proposed approach.

In this paper, we identify a new security issue, called the malicious energy attack, in sustainable wireless communication networks (SWCNs). We show that by providing extra energy to specific nodes, a malicious energy source (MES) can intentionally manipulate the routing path of SWCNs. The efficiency of energy attack depends on which nodes to be attacked. To enhance the efficiency of energy attack, a reinforcement learning technique, Q-Learning, is used to develop an intelligent energy attack (Q-IEA) policy for MES. Through interacting with the network environment, the Q-IEA can intelligently take attack actions without having to know the details of the routing method at the network layer. This function can greatly enhance the adaptability of MES to different routing protocols and network topologies. Simulation results verify that Q-IEA can significantly manipulate the routing path of the targeted traffic on demand.

Increasing bandwidth requirements for home and business Internet connections as well as fiber links for 5G wireless drive the need for higher data rate passive optical networks (PON). PON is a cost-effective way to deploy fiber. A point-to-multipoint topology with passive optical splitters is used to connect many subscribers to one central node. The next development step in PON is 50 Gbit/s per wavelength, which requires improvements in error correction coding and receiver signal processing. This paper investigates digital equalization for PON with respect to performance with soft decision error correction decoding.

In this paper, we investigate a computing task offloading problem in a cloud-based autonomous vehicular network (C-AVN), from the perspective of long-term network wide computation load balancing. To capture the task computation load dynamics over time, we describe the problem as an Markov decision process (MDP) with constraints. Specifically, the objective is to minimize the expectation of a long-term total cost for imbalanced base station (BS) computation load and task offloading decision switching, with per-slot computation capacity and offloading latency constraints. To deal with the unknown state transition probability and large state-action spaces, a multi-agent deep Q-learning (MA-DQL) module is designed, in which all the agents cooperatively learn a joint optimal task offloading policy by training individual deep Q-network (DQN) parameters based on local observations. To stabilize the learning performance, a fingerprint-based method is adopted to describe the observation of each agent by including an abstraction of every other agent’s updated state and policy. Simulation results show the effectiveness of the proposed task offloading framework in achieving long-term computation load balancing with controlled offloading switching times and per-slot QoS guarantee.

Nodes in wireless Internet of Things (IoT) sensor networks are heterogeneous in nature. This heterogeneity can come from energy and security resources available at the node level. Besides, these resources are usually limited. Efficient cluster head (CH) selection in rounds is the key to preserving energy resources of sensor nodes. However, energy and security resources are contradictory to one another. Therefore, it is challenging to ensure CH selection with appropriate security resources without decreasing energy efficiency. Coverage and energy optimization subject to a required security level can form a solution to the aforementioned trade-off. This paper proposes a security level aware CH selection algorithm in wireless sensor networks for IoT. The proposed method considers energy and security level updates for nodes and coverage provided by associated CHs. The proposed method performs CH selection in rounds and in a centralized parallel processing way, making it applicable to the IoT scenario. The proposed algorithm is compared to existing traditional and emerging CH selection algorithms that apply security mechanisms in terms of energy and security efficiencies.

In this paper, we propose an online deep learning (DL)-based channel estimation (CE) for massive multiple-input multiple-output (MIMO) systems with limited pilots, where the training stage can be implemented online based on real-time received pilot signals. This is realized by introducing a sparsifying loss function that is model-free and does not need ground truth labeled channel data. Simulation results show that the proposed online DL-based scheme achieves comparable channel estimation performance to traditional model-based compressive sensing (CS) algorithms while enjoying a much faster computation during the channel inferencing stage. In addition, the proposed scheme is also robust to various model mismatches and is able to track the change of the underlying propagation environment.

Cellular networks have been successfully evolved over the decades. Especially, Long-Term Evolution (LTE) has been exceedingly successful and the security threats against LTE systems have increased rapidly. Particularly, tracking LTE user devices has been shown to be effective as the temporary user identifiers (IDs), which are used in LTE systems to indicate LTE user devices in the system, are easily extracted and used to locate targeted devices by passive eavesdroppers. In this paper, we investigate the impact of uplink power control on the probability of successful user tracking by an adversary whose location is unknown. We devise the notion of average inference error probability in order to measure the level of users’ location privacy. Moreover, we derive the closed-form expression of the approximated average inference error probability and formulate an optimization problem for maximizing the average inference error probability under a constraint of an allowable power budget for each user. For defense, we propose a power control scheme able to effectively degrade an adversary’s inference ability by 50% when 10 users are scheduled in each transmission time slot, which will result in almost 100% inference error at the adversary over multiple time slots.

Leveraging powerful deep learning techniques, the end-to-end (E2E) learning of communication system is able to outperform the classical communication system. Unfortunately, this communication system cannot be trained by deep learning without known channel. To deal with this problem, a generative adversarial network (GAN) based training scheme has been recently proposed to imitate the real channel. However, the gradient vanishing and overfitting problems of GAN will result in the serious performance degradation of E2E learning of communication system. To mitigate these two problems, this paper proposes a residual aided GAN (RA-GAN) based training scheme. Particularly, inspired by the idea of residual learning, we propose a residual generator to mitigate the gradient vanishing problem by realizing a more robust gradient back propagation. Moreover, to cope with the overfitting problem, we reconstruct the loss function for training by adding a regularizer, which limits the representation ability of RA-GAN. Simulation results show that the trained residual generator has better generation performance than conventional generator, and the proposed RA-GAN based training scheme can achieve the near-optimal block error rate (BLER) performance.

Understanding the required quality of service (QoS) for guaranteeing the necessary video quality for Internet video streaming services is important for enabling network operators to provide high-quality networks. Recent trace-based approaches that infer video quality through machine-learning based traffic-feature analysis have a drawback in terms of the maintenance cost for updating their analysis model due to the massive number of videos uploaded daily. In this paper, we construct a QoS analysis model that calculates the required throughput for delivering video with a certain resolution by using encoding information instead of traffic tracing. For accurate modeling, we consider and formulate communication overhead, i.e. streaming behavior, retransmission, and headers. Through experiments, we demonstrate that our model can perform in two use cases: (1) estimating effective resolution for arbitrary QoS, and (2) calculating the required QoS for guaranteeing the necessary video quality.

It is well established that 5G will impact not only the end-users by allowing several new services, but also the vertical industry and network operators business. 5G will open the business market to new stakeholders with the introduction of Network Slicing, namely the vertical or tenant, the network slice provider, and the infrastructure provider. The Network Slice provider sells end-to-end network slices (virtual end-to- end mobile network) to the vertical while leasing virtual and physical resources from Infrastructure Providers to enforce these end-to-end network slices. Accordingly, there is a need to establish Service Level Agreement (SLA) among these actors to ensure: (1) that the service is well-delivered to the vertical and (2) the infrastructure providers are respecting their involvement with the network slice provider. To fill this gap, in this paper, we propose a trust architecture to automatically manage the SLAs and apply penalties and compensations if the SLAs are not respected by one of the involved actors.

Knowledge sharing in IoV shows great potential for future vehicular networks. Vehicles, platoons and even traffic infrastructures can exchange the driving experiences or sensing data to facilitate intelligent transportation applications such as autodriving and traffic analysis. However, it is challenging for vehicular knowledge-sharing systems to address the issues brought by information security and vehicular mobility. Although blockchain technology shows defensibility in dealing with trust issues, it is difficult to be applied in large-scale vehicular networks due to the computation consumption of mining process and frequent synchronization of ledger. In this paper, we propose a directed acyclic graph (DAG) enabled knowledge-sharing framework in which vehicular knowledge is encapsulated as a site in the DAG. A new tip selection algorithm (TSA) and a fast authentication scheme for cross-regional vehicles are designed to reduce computation and storage expenditure. Simulation results show that the proposed DAG framework can achieve a higher knowledge sharing quality and lower authentication latency compared with traditional DAG systems.

Benefiting from the breakthrough development of the fifth generation (5G), beyond 5G (B5G) wireless communication networks and Artificial Intelligence (AI) in recent years, the artificial intelligence of things (AIoT) is a new trend in the future. AIoT devices often have high-quality wireless video transmission requirements. However, the propagating signals at millimeter wave suffer from high propagation loss and sensitivity to blockage, resulting in the received video is vulnerable to be interfered. Due to the ability of Deep Learning (DL) to discover and learn good representations, some DL methods have achieved breakthrough performance in video recovery. However, most of these methods cannot exploit information from the higher to lower level to refine themselves. In this paper, we propose a novel retrospective thinking based multi-agent (ReTMA) system to solve the interference problem experienced on wireless channels. Compared with other plain feedback models, we add a retrospective agent on the feedback loop, which makes the entire system have stronger capabilities to learn good representative features. We further formulate it as a Stackelberg game to analyze the dependency relationship between the agents and facilitate the complex training issue of the multiple agents. To verify the feasibility of ReTMA system, we randomly add masks to simulate the severe interference received by the video frames in wireless transmissions. Experimental results show that the performances of similarity index measure (SSIM), peak signal-to-noise ratio (PSNR) and classification accuracy all achieve significant gains compared with those of other plain feedback models at different mask ratios.

Resource allocation overbooking is an approach used by cloud providers that allocates more virtual resources than available on physical hardware, which may imply service quality degradation. Docker in cloud computing environments is being increasingly used due to their fast provisioning and deployment, while the impact of overbooking of resources allocation due to multi-tenancy remains overlooked. This paper proposes a machine learning model to detect overbooking in Kubernetes environments within the docker container. The proposed model continuously monitors distributed container OS usage and application performance metrics. The collected metrics are used as input to a machine learning model that identifies multi-tenancy interference incurring in application performance degradation. Experiments performed on a Kubernetes cluster with a Docker-based Big Data processing application showed that our proposed model could detect resource overbooking with up to 98% accuracy. This implies an overbooking on a resource of up to 1.2 in the client’s domain.

Federated learning-based navigation has received much attention in vehicular IoT. The intention is to employ a big number of end-users for data collection along different trajectories and perform local training of a global learning model to substitute the global positioning system (GPS) in urban areas. The prerequisites for its commercialization, however, lie in the location-dependent input data trustworthiness and participants’ privacy preservation. In this paper, we propose a privacy-preserving proof-of-location mechanism using blockchain to meet these conditions. Specifically, the proposed scheme utilizes a Threshold Identity-Based Encryption (TIBE) system for the generation of secret shares, such that each anonymous location proof can only be verified with at least a threshold number of participants. In addition, the proposed scheme exploits a cuckoo filter for the secure and efficient maintenance and dissemination of location proofs. Systematic security analysis is conducted to demonstrate the fulfillment of harsh security requirements. Performance evaluations are carried out to validate the computation efficiency in comparison with an oblivious transfer (OT) protocol, which has been widely adopted for secure data acquisition.

In this paper, we investigate a deep learning architecture for lightweight online implementation of a reconfigurable intelligent surface (RIS)-aided multi-user mobile edge computing (MEC) system, where the optimized performance can be achieved based on user equipment’s (UEs’) location-only information. Assuming that each UE is endowed with a limited energy budget, we aim at maximizing the total completed task-input bits (TCTB) of all UEs within a given time slot, through jointly optimizing the RIS reflecting coefficients, the receive beamforming vectors, and UEs’ energy partition strategies for local computing and computation offloading. Due to the coupled optimization variables, a three-step block coordinate descending (BCD) algorithm is first proposed to effectively solve the formulated TCTB maximization problem iteratively with guaranteed convergence. The location-only deep learning architecture is then constructed to emulate the proposed BCD optimization algorithm, through which the pilot channel estimation and feedback can be removed for online implementation with low complexity. The simulation results reveal a close match between the performance of the BCD optimization algorithm and the location-only data-driven architecture, all with superior performance to existing benchmarks.

The differential pulse position modulation (DPPM) is one of the popular power-efficient schemes for supporting visible light communication in underwater environments and across the air-water interface. Despite such an advantage, DPPM does not efficiently utilize the available channel capacity. This paper aspires to tackle such shortcomings by striking a better balance between power and bandwidth efficiency. Particularly, L-DPPM is considered where a block of M input data bits is mapped into one of the L distinct waveforms containing only one ‘on’ chip. A novel encoding algorithm and frame structure are proposed in order to shorten the time between consecutive symbols and consequently improve the bit rate of L-DPPM. The idea is based on avoiding bit patterns that contribute the most to bandwidth inefficiency. The proposed algorithm explores a number of bit patterns remapping through simple complement and shifting operations. A detailed frame structure with all necessary control bits is provided. Overall, boosting the bandwidth efficacy comes at the expense of a slight increase in control bit count and transmission power. The simulation results demonstrate the effectiveness of the proposed encoding algorithm and provide guidelines for determining M for best performance.

Throughput-Outage scaling laws for single-hop cache-aided device-to-device (D2D) communications have been extensively investigated under the assumption of the protocol model. However, the corresponding performance under physical models has not been explored; in particular it remains unclear whether link-level power control and scheduling can improve the asymptotic performance. This paper thus investigates the asymptotic throughput-outage tradeoff and derives its outer bound for cache-aided D2D networks under two common physical models. The results show that the asymptotic performance of the network under physical models is identical to that under the protocol model when requests are served with equal quality. This indicates that the throughput-outage performance cannot be improved asymptotically by using link-level power control and scheduling.

The ability to accurately forecast network resource utilization is vital in next-generation wireless networks. Based on the predicted load, telecom operators can proactively allocate network resources in an efficient way. In this paper, we perform a thorough analysis of a cellular network downlink load dataset collected at millisecond resolution. We first evaluate various statistical metrics of the physical resource block (PRB) utilization data to investigate its predictability. Then, we develop deep learning-based models to forecast PRB utilization in radio access networks (RANs). In particular, we propose univariate and multivariate long short-term memory (LSTM) network-based architectures for the forecasting task and investigate the impact of various prediction horizons and history lengths. When predicting PRB utilization, our approach showed up to 49% improvement in the Coefficient of Determination (r2 score) and 19.5% decrease in the Root Mean Square Error (RMSE) compared with the baseline methods used.

Demand side management (DSM) presents an effective tool to regulate customers’ energy consumption. However, DSM requires customer-utility interaction based on fine-grained data, which jeopardizes the privacy of the customers. To overcome such a threat, several proposals are made based on noise addition, encryption techniques, or extra hardware. Unfortunately, these solutions either impact the integrity of the data, incur additional computational complexity, or add extra cost. Recently, blockchain has been adopted to support smart grid applications while promoting customer’s anonymity. However, special measures are yet to be taken to support both customer’s anonymity and data unlinkability, hence preserving the customer’s privacy. This paper achieves this goal by proposing a novel networking strategy based on a private blockchain. The proposed strategy employs a group signature to ensure the customer’s anonymity and data unlinkability. Further, a novel decentralized random number generation scheme is proposed to support customer-utility interaction for DSM, billing, and auditing while ensuring data unlinkability. We present an implementation of the proposed blockchain-based strategy, investigate its scalability, and provide low-bound empirical expressions on computational time and storage overhead.