In this work, we develop a framework that jointly decides on the optimal location of wireless extenders11Wireless extenders are wireless devices used to extend the coverage area of a main access point; they are also called mesh points, since the main access point refers to a mesh gateway in a WMN. and the channel configuration of wireless extenders and access points (APs) in a Wireless Mesh Network (WMN). Artificial Intelligence (AI) is adopted to support network autonomy and to capture insights on system and environment evolution. We propose a Self-X (self-optimizing and self-learning) framework that encapsulates both environment and intelligent agent to reach optimal operation through sensing, perception, reasoning and learning, in a truly autonomous fashion. The agent derives adequate knowledge from previous actions, improving the quality of future decisions. Extensive simulations are run to validate its fast convergence, improved throughput and resilience to dynamic interference conditions.

In this paper, we present our recent studies on pilot allocation and interference coordination for heterogeneous networks (HetNets) using massive multi-input-multi-output (mMIMO) technology in time division duplex (TDD) mode, where the macro-cell base station (MBS) and overlaid small-cell base stations (SBSs) share the same time-frequency resources, and thus co-tier and cross-tier interferences occur. We investigate how to coordinate such undesirable interference for maximizing the system sum-rate under required constraints. As the first approach, we present an uplink pilot allocation scheme to enhance the downlink system sum-rate by coordinating downlink cross-tier interference to small-cell users (SUs) based on the estimated uplink channel state information. In this approach, we study the optimum pilot allocation against the trade-off between two degradation factors, i.e., uplink pilot overhead and downlink cross-tier interference. As the second approach, we present a dynamic SBS clustering scheme to mitigate dominant co-tier interference among small cells (SCs), where the SBS clustering is carried out based on potential mutual co-tier interference strength between two SCs. We also consider the SBS precoder design in each cluster to further improve the downlink sum rate of SCs under SBS power constraint. Simulation results show that our proposals are useful approaches to enhance the sum rate performance in TDD-mMIMO-HetNets.

In this paper, we investigate the bandwidth and power allocation problem in remote radio head cluster (RRHC)-based millimeter wave (mm-wave) massive MIMO heterogeneous cloud radio access networks with limited fronthaul capacity. The coordinated multipoint transmission is applied in each RRHC for cancelling the intra-cluster interference. To avoid the inter-tier interference, distinct bandwidths are allocated to macro base station and RRHs. Following this, we formulate a bandwidth and power allocation optimization problem to maximize the downlink weighted sum rate of the system subject to per-RRHC power and fronthaul capacity constraints, which is a non-convex optimization problem and is difficult to directly solve. Next, we fix the bandwidth allocation and the original problem can be divided into two independent optimization problems, i.e., the weighted sum rate maximization problems of MUs and RRH users, respectively. For the former, the convex optimization technique can be used to solve it. As for the latter, a two-loop iterative algorithm is proposed to deal with it. Specifically, we propose the price-based outer iteration to control the fronthaul capacity and the weighted minimum mean square error-based inner iteration to obtain the power allocation. To this end, a 1-D search method is adopted to find the optimal bandwidth allocation. Finally, numerical results are conducted to verify the effectiveness of the proposed algorithms under different parameters.

Algorithms for distributed coordination and control are increasingly being used in smart grid applications including peer-to-peer energy trading and sharing to improve reliability and efficiency of the power system. However, for realistic deployment of these algorithms, their designs should take into account the suboptimal conditions of the communication network, in particular the communication links that connect the energy trading entities in the energy network. This study proposes a distributed adaptive primal (DAP) routing algorithm to facilitate communication and coordination among proactive prosumers in an energy network over imperfect communication links. The proposed technique employs a multi-commodity flow optimization scheme in its formulation with the objective to minimize both the communication delay and loss of energy transactional messages due to suboptimal network conditions. Taking into account realistic constraints relating to network delay and communication link capacity between the peers, the DAP routing algorithm is used to evaluate network performance using various figures of merit such as probability of signal loss, message delay, congestion and different network topologies. Further, we address the link communication delay problem by redirecting traffic from congested links to less utilized ones. The results show that the proposed routing algorithm is robust to packet loss on the communication links with a 20% reduction in delay compared with hop-by-hop adaptive link state routing algorithm.

This paper discusses technology challenges and opportunities to embrace artificial intelligence (AI) era in the design of wireless networks. We aim to provide readers with motivation and general methodology for adoption of AI in the context of next-generation networks. First, we discuss the rise of network intelligence and then, we introduce a brief overview of AI with machine learning (ML) and their relationship to self-organization designs. Finally, we discuss design of intelligent agent and it's functions to enable knowledge-driven wireless networks with AI.

Wireless mesh networks (WMNs) have been extensively studied for nearly two decades as one of the most promising candidates expected to power the high-bandwidth, high-coverage wireless networks of the future. However, consumer demand for such networks has only recently caught up, rendering efforts at optimizing WMNs to support high capacities and offer high QoS, while being secure and fault-tolerant, more important than ever. To this end, a recent trend has been the application of machine learning (ML) to solve various design and management tasks related to WMNs. In this work, key ML techniques are discussed and past efforts applying them in WMNs are analyzed, while noting some existing issues and suggesting potential solutions. Directions are provided on how ML could advance future research. Recent developments in the field are also examined.

In this paper, we investigate asynchronous coded caching in fog radio access networks (F-RAN). To minimize the fronthaul load, the encoding set collapsing rule and encoding set partition method are proposed to establish the relationship between the coded-multicasting contents in asynchronous and synchronous coded caching. Furthermore, a decentralized asynchronous coded caching scheme is proposed, which provides asynchronous and synchronous transmission methods for different delay requirements. The simulation results show that our proposed scheme creates considerable coded-multicasting opportunities in asynchronous request scenarios.

The IEEE 1901 powerline standard can be deployed using orthogonal frequency division multiplexing (OFDM) since it is robust over impulsive channels. However, the powerline channel picks up impulsive interference that the conventional OFDM driver cannot combat. Since the probability density function (PDF) of OFDM amplitudes follow the Rayleigh distribution, it becomes difficult to correctly predict the existence of impulsive noise (IN) in powerline systems. In this study, we use companding transforms to convert the PDF of the conventional OFDM system to a uniform distribution which avails the identification and mitigation of IN. Results show significant improvement in the output signal-to-noise ratio (SNR) when nonlinear optimization search is applied. We also show that the conventional PDF leads to false IN detection which diminishes the output SNR when nonlinear memoryless mitigation scheme such as clipping or blanking is applied. Thus, companding OFDM signals before transmission helps to correctly predict the optimal blanking or clipping threshold which in turn improves the output SNR performance.

As the age of the Internet of Things (IoT) continues to flourish, the concept of smart healthcare has taken an unprecedented turn due to interdisciplinary thrusts. To carry the big healthcare data ema nating from the plethora of bi 0-sens ors and machines in the IoT sensing plane to the central cloud, next generation high-speed delivery networks are essential. On the other hand, once the IoT data are delivered to the cloud, the massive IoT healthcare data are processed and analyzed em-ploying the state-of-the-art analytics tools such as deep machine learning and so forth. However, given the explosion of big data (from various sources in addition to the healthcare data), the delivery network as well the cloud may experience network and computational congestion, respectively. This may impact the realtime analytics of the healthcare data, e.g., critical for in-house patients and senior citizens aging at home. To address this issue, the emerging IoT edge analytics concept can be regarded as a promising solution to process the big healthcare data close to the source. For larg e-s cale IoT dep loym ents, this fu nctio nality is critical because of the sheer volumes of Data being generated. In this paper, we propose a deep learning based IoT edge analytics approach to support intelligent healthcare for residential users. The performance of the proposal is validated using computer-based simulation for online training of a real dataset. The reported results of our proposal exhibit encouraging performance in terms of low loss rate, high accuracy, and low execution time to support near real-time actionable decision making on the healthcare data.

In this paper, we investigate pilot allocation problem in two-tier time division duplex (TDD) heterogeneous network (HetNet) with mMIMO. First, we propose a new pilot allocation scheme to maximize ergodic downlink sum rate of macro users (MUs) and small cell users (SUs), where the uplink pilot overhead and cross-tier interference are jointly considered. Then, we theoretically analyze the formulated problem and propose a low complexity one-dimensional search algorithm to obtain the optimum pilot allocation. In addition, we propose two suboptimal pilot allocation algorithms to simplify the computational process and improve SUs' fairness, respectively. Finally, simulation results show that the performance of the proposed scheme outperforms that of the traditional schemes.

In power line communication (PLC) systems using multi-carrier modulation such as orthogonal frequency division multiplexing (OFDM), reduction of peak-to-average power ratio (PAPR) of OFDM signal and mitigation of the influence of impulsive noise are challenging problems. To deal with these problems in PLC-OFDM systems, this paper proposes an enhanced selected mapping (e-SLM) technique that jointly achieves PAPR reduction on the transmitter side and impulsive noise detection accuracy improvement on the receiver side, where the transmitter selects the best transmit sequence among possible candidates so as to minimize the weighted sum of transmit signal peak power and the estimated receive one, provided that channel state information is available on the transmitter side. On the receiver side, a nonlinear filter is applied to suppress impulsive noise, where the received signal amplitudes exceeding a given threshold are blanked as impulsive noise. Using the proposed e-SLM, peak power of the received signal is reduced below a given threshold level, so that impulsive noise is effectively detected with high accuracy. We evaluate bit error rate (BER) performance of PLC-OFDM systems using the proposed e-SLM method under PLC channel in presence of impulsive noise. Simulation results show the effectiveness of the proposed method in OFDM PLC systems.

In this work, we develop a framework that jointly decides on the optimal location of wireless extenders and the channel configuration of extenders and access points (APs) in a Wireless Mesh Network (WMN). Typically, the rule-based approaches in the literature result in limited exploration while reinforcement learning based approaches result in slow convergence. Therefore, Artificial Intelligence (AI) is adopted to support network autonomy and to capture insights on system and environment evolution. We propose a Self-X (self-optimizing and self-learning) framework that encapsulates both environment and intelligent agent to reach optimal operation through sensing, perception, reasoning and learning in a truly autonomous fashion. The agent derives adequate knowledge from previous actions improving the quality of future decisions. Domain experience was provided to guide the agent while exploring and exploiting the set of possible actions in the environment. Thus, it guarantees a low-cost learning and achieves a near-optimal network configuration addressing the non-deterministic polynomial-time hardness (NP-hard) problem of joint channel assignment and location optimization in WMNs. Extensive simulations are run to validate its fast convergence, high throughput and resilience to dynamic interference conditions. We deploy the framework on off-the-shelf wireless devices to enable autonomous self-optimization and self-deployment, using APs and wireless extenders.

In this paper, we propose a self-deployment approach for finding the optimal placement of extenders in which both the wireless back-haul and front-haul throughput of the extender are optimized. We present an artificial intelligence (AI) case based reasoning (CBR) framework that enables autonomous self-deployment in which the network can learn the environment by means of sensing and perception. New actions, i.e. extender positions, are created by problem-specific optimization and semi-supervised learning algorithms that balance exploration and exploitation of the search space. An IEEE 802.11 standard compliant simulations are performed to evaluate the framework on a large scale and compare its performance against existing conventional coverage maximization approaches. Experimental evaluation is also performed in an enterprise environment to demonstrate the competence of the proposed AI-framework in perceiving such a dense scenario and reason the extender deployment that achieves user quality of service (QoS). Throughput fairness and ubiquitous QoS satisfaction are achieved which provide a leap to apply AI-driven self-deployment in wireless networks.

With the advancements of next-generation programmable networks, traditional rule-based decision- making may not be able to adapt effectively to changing network and customer requirements and provide an optimal customer experience. CEM components and implementation challenges with respect to operator, network, and business requirements must be understood to meet required demands. This article gives an overview of CEM components and their design challenges. We elaborate on data analytics and artificial intelligence driven CEM and their functional differences. This overview provides a path toward an autonomous CEM framework in next-generation networks and sets the groundwork for future enhancements.

Future indoor networks need to provide broadband services either by increasing bandwidth, using higher order modulations and high spatial reuse. To this end, ultra-dense Wi-Fi networks need to be deployed. This work discusses the challenges of ultra-dense Wi-Fi networks, where inter-AP distances are 5 m or less, and discuss necessary requirements to maximize the high spatial reuse of radio resources through the use of adaptive Clear Channel Assessment (CCA) thresholds and downward facing directional antennas. We also examine the significant impact of uplink transmissions, and present the necessary uplink transmit power control in ultra-dense deployments to avoid the interference coming from hidden nodes in neighboring APs. With the proposed techniques and antennas in place, we show that the limitations placed by high contention and interference in ultra-dense deployments can be overcome to provide over 8 χ gains in downlink throughput when compared with APs with omni-directional and unadjusted CCA thresholds.