Search

Abstract The active distribution network (ADN) can obtain measurement data, estimate system states, and control distributed energy resources (DERs) and flexible loads to ensure voltage stability. However, the ADN is more vulnerable to cyber attacks due to the recent wave of digitization and automation efforts. In this article, false data injection (FDI) attacks are focused on and they are classified into two types, that is, type I attacks on measurement data and type II attacks on control commands. After studying the impact of these two FDI attacks on the ADN, a new threat is revealed called coordinated FDI attack, which can maximize the voltage deviation by coordinating type I and type II FDI attacks. From the attacker's perspective, the scheme of CFDI is proposed and an algorithm is developed to find the optimal attack strategy. The feasibility of CFDI attacks has been validated on a smart distribution testbed. Moreover, simulation results on an ADN benchmark have demonstrated that CFDI attacks could cause remarkable voltage deviation that may deteriorate the stability of the distribution network. Moreover, the impact of CFDI attacks is higher than pure type I or type II attacks. To mitigate the threat, some countermeasures against CFDI attacks are also proposed.

Deep learning has been used for optimizing a multitude of wireless problems. Yet most existing works assume that training and test samples are drawn from the same distribution, which is not true in dynamic environments. As a result, a deep neural network (DNN) may require retraining with newly collected samples to adapt to a new scenario. The retraining complexity can be reduced by introducing inductive biases, which can either be learned automatically or be embedded in DNNs manually. For example, inductive biases can be learned from related tasks by meta-learning in the form of good initializations, reusable modules and feature extractors, or be introduced by designing structural DNNs such as graph neural networks (GNNs) with appropriate permutation equivariance (PE) properties. All previous works on meta-learning overlooked the prior-known PE properties, which widely exist in wireless policies and can be leveraged to improve sample efficiency. This article reviews meta-learning for wireless communications and compares it with GNNs trained for a single task. We first introduce meta-learning by comparing it with conventional learning, analyze its inductive biases, and review its applications for radio transmission. We then compare meta-learning with GNNs trained for a single task by taking beamforming problem as an example to show their pros and cons in solving the mismatch issues. Simulation results show that GNNs are sample efficient while meta-learning can reduce the time complexity for adaptation. Both the considered GNNs and meta-learning methods are inefficient for adapting to environments with time-varying problem scales.

Abstract Background Postmenopausal osteoporosis (PMOP) is a common condition that leads to a loss of bone density and an increased risk of fractures in women. Recent evidence suggests that exosomal miRNAs are involved in regulating bone development and osteogenesis. However, exosomal miRNAs as biomarkers for PMOP diagnosis have not been systematically evaluated. In this study, we aim to identify PMOP-associated circulating exosomal miRNAs and evaluate their diagnostic performance. Methods We performed next-generation sequencing and bioinformatics analysis of plasma exosomal miRNAs from 12 PMOP patients and 12 non-osteoporosis controls to identify PMOP-associated exosomal miRNAs, and then validated them in an independent natural community cohort with 26 PMOP patients and 21 non-osteoporosis controls. Exosomes were isolated with the size exclusion chromatography method from the plasma of elder postmenopausal women. The plasma exosomal miRNA profiles were characterized in PMOP paired with controls with next-generation sequencing. Potential plasma exosomal miRNAs were validated by qRT-PCR in the validation cohort, and their performance in diagnosing PMOP was systematically evaluated with the receiver operating characteristic curve. Results Twenty-seven miRNAs were identified as differentially expressed in PMOP versus controls in sequencing data, of which six exosomal miRNAs (miR-196-5p, miR-224-5p, miR320d, miR-34a-5p, miR-9-5p, and miR-98-5p) were confirmed to be differentially expressed in PMOP patients by qRT-PCR in the validation cohort. The three miRNAs combination (miR-34a-5p + miR-9-5p + miR-98-5p) demonstrated the best diagnostic performance, with an AUC = 0.734. In addition, the number of pregnancies was found to be an independent risk factor that can improve the performance of exosomal miRNAs in diagnosing PMOP. Conclusions These results suggested that the plasma exosomal miRNAs had the potential to serve as noninvasive diagnostic biomarkers for PMOP.

In recent years, due to the extensive application and inherent fire hazard of cable materials, the combustion characteristics of frequently used cables, including electrical cables, wires, optical fibers, and network cables have been studied based on ISO 5660 cone calorimetry. The fire hazard associated with these cables under different radiation intensities was explored in this study, with parameters such as time to ignition (TTI), heat release rate per unit area (HRRPUA), peak heat release rate (PHRR), total heat release (THR), and mass loss rate (MLR) being investigated for each cable type. Based on an experimental analysis, the risk of fire for all four cable types was augmented by an increase in the external radiation intensity, with electrical cables considered as posing the greatest risk. Regarding smoke toxicity, the lowest risk of smoke toxicity was demonstrated by the network cable, with an FED (fractional effective dose) of 0.0203, followed by optical fibers, with an FED of 0.0507; electrical wires, with an FED of 0.0417; and electrical cables, with an FED of 0.0501. Notably, no significant distinctions were exhibited by the other three cable types, and the smoke toxicity of all four cables did not reach lethal concentration levels in humans. Consequently, considering both thermal hazard and smoke toxicity, it became evident that electrical cables posed the greatest overall fire hazard.

Dual connectivity has been regarded as an effective mechanism for mobility management to reduce the handover latency and improve transmission reliability. In this article, we investigate a dual connectivity-aided proactive mobility management scheme for mobile users requesting realtime service with minimum data rate requirements. The proactive mobility management consists of a predictive resource reservation scheme and a proactive handover scheme. The predictive resource reservation scheme is optimized to improve the network throughput or user fairness with the predicted average channel gains and the predicted set of connected base stations for each user. The proactive handover scheme aims to avoid frequent and delayed handovers, and reduce the computational complexity by partitioning cell-center and cell-edge users and choosing different numbers of connected base stations for the two types of users. Simulation results show that the proposed scheme outperforms reactive counterparts, which can effectively reduce service outages meanwhile support higher network throughput when the traffic load is heavy, or improve user fairness meanwhile cause less service outages when the traffic load is light.

A distributed interference coordination strategy based on multi-agent deep reinforcement learning was investigated to meet the requirements of file downloading traffic in interference networks.By the proposed strategy transmission scheme could be adjusted adaptively based on the interference environment and traffic requirements with limited amount of information exchanged among nodes.Simulation results show that the user satisfaction loss of the proposed strategy from the optimal strategy with perfect future information does not exceed 11% for arbitrary number of users and traffic requirements.

Proactive caching at the base station (BS) is a promising way to leverage the user-behavior-related information to boost network throughput and improve user experience. However, the gain of caching at the mobile edge highly depends on random user behavior and is largely compromised by the uncertainty in predicting behavior-related information. First, the local file popularity in each cell may not be skewed. Second, the local file popularity varies quickly due to user mobility even if the lifetime of each file is long. Furthermore, considering the small population of users that initiate requests in each cell, the local popularity in the next cache update period is not easy to predict accurately, because users may not request their interested files in this period, despite that the popularity can be indirectly obtained by predicting the mobility and preference of each individual user in a cell. To address such issue, in this paper, we integrate recommendation with caching at BS, aiming at improving cache efficiency whereas not violating user preference. In particular, we propose a temporal-spatial recommendation policy, which can guide mobile users to request their preferred files in proper time and place, so as to make local popularity peakier. We do not assume that the user preference, the impact of the recommendation on request probability, and the mobility pattern are known. Hence, we resort to deep reinforcement learning to optimize recommendation and caching policy. To deal with the difficulty in predicting local popularity in the next cache replacement period, we model the user preference and request probability with Bernoulli mixture distribution and hence can estimate them separately. The simulation results demonstrate that the proposed policy can reduce the cache miss number, compared to the policies without any recommendation and without temporal-spatial recommendation.

The gain from proactive caching at mobile devices highly relies on the accurate prediction of user demands and mobility, which, however, is hard to achieve due to the random user behavior. In this paper, we leverage personalized content recommendation to reduce the uncertainty of user demands in sending requests. We formulate a joint content pushing and recommendation problem that maximizes the net profit of a mobile network operator. To cope with the challenges in modeling and learning user behavior, we establish a reinforcement learning (RL) framework to resolve the problem. To circumvent the curse of dimensionality of reinforcement learning for the joint problem, that is, with very large action and state spaces, we decompose the original problem into two RL problems, where two agents with different goals operate together, and we limit the number of possible actions in each state of the pushing agent by harnessing the well-learned recommendation policy. To enable the generalization of action values from experienced states to the unexperienced states with function approximation, we find a proper way to represent the state and action of the pushing agent. Then, we resort to double deep-Q network with dueling architecture to solve the two problems. The simulation results show that the learned recommendation and pushing policies are able to converge and can increase the net profit significantly compared with baseline policies.

Caching at wireless edge is a promising way to satisfy the explosively increasing mobile data demands, if future content popularity is known in advance. However, the time-varying nature of content popularity makes the popularity prediction far from perfect, which inevitably degrades the gain from caching. In this paper, we resort to a hybrid proactive and reactive policy to deal with the dynamics of popularity, in particular the hybrid of proactive probabilistic caching policy and least recently used policy, which are appropriate respectively for the contents with low and high dynamic popularity. We divide the contents requested in a region into two classes, where one can be modeled by independent reference model (IRM) and the other can be modeled by shot noise model (SNM). To maximize the total successful offloading ratio achieved by caching the two classes of contents, we optimize the hybrid caching policy including both the cache resource allocation to each class of contents and the probability of caching each IRM content. We find the optimal solution to the problem for general case, and provide a closed-form solution in a special case to gain insights. To provide a viable solution for practical use, we propose a heuristic method to obtain the optimal allocation fraction, and predict the popularity distribution and the allocation fraction using neural networks with historical data. We validate our analytical results by simulation results via synthetic datasets. We evaluate the performance of the proposed hybrid caching policy via synthetic data generated by SNM and two real datasets, and compare it with the proactive policy, the reactive policy and the existing hybrid proactive and reactive policy.

Future average channel gains are recently reported predictable within a minute-level horizon for a mobile user. Predictive resource allocation for non-realtime service with future average channel gains in a time window of one or more minutes long has been demonstrated effective in improving user experience and network throughput as well as reducing energy consumption. Nonetheless, existing studies for predictive resource allocation never consider inter-cell interference (ICI), which severely limits the user experience and network performance in densely-deployed cellular networks. This paper investigates predictive resource allocation in heterogeneous networks, where some base stations generate strong ICIs to some mobile users requesting non-realtime service. Optimizing the resource allocation in a predictive manner for interference networks is challenging, since how to allocate future resources depends on the future signal to interferenceplus-noise ratio, which in turn relies on the assigned resources. To deal with this difficulty, we introduce a predictive interference coordination scheme to divide all BSs and users into groups, where the BS-user pairs in each group can communicate simultaneously in a second-level frame. Then, we optimize a common resource allocation plan for all users in each group. The plan essentially determines which BSs in the group should be muted and the users should be associated with which active BS in each frame of the prediction window. By resorting to graph theory, we obtain the optimal solution and derive a low-complexity algorithm. Simulation results show that the proposed scheme outperforms existing relevant methods in terms of user satisfactory rate in heterogeneous networks with heavy traffic load.

Network slicing is born as an emerging business to operators by allowing them to sell the customized slices to various tenants at different prices. In order to provide better-performing and costefficient services, network slicing involves challenging technical issues and urgently looks forward to intelligent innovations to make the resource management consistent with users' activities per slice. In that regard, deep reinforcement learning (DRL), which focuses on how to interact with the environment by trying alternative actions and reinforcing the tendency actions producing more rewarding consequences, is assumed to be a promising solution. In this paper, after briefly reviewing the fundamental concepts of DRL, we investigate the application of DRL in solving some typical resource management for network slicing scenarios, which include radio resource slicing and priority-based core network slicing, and demonstrate the advantage of DRL over several competing schemes through extensive simulations. Finally, we also discuss the possible challenges to apply DRL in network slicing from a general perspective.

Machine learning is a powerful tool to predict user behavior and harness the vast amount of data measured in cellular networks. Predictive resource allocation is a promising approach to take advantage of the prediction for the mobility and traffic load related user behavior. This paper strives to boost the performance of under-utilized networks by predicting behavior-related information from historical data with deep learning. We first propose a hierarchical and multi-timescale radio resource management scheme for non-realtime service that only needs coarse-grained future knowledge, by taking multi-input-multi-output orthogonal frequency multi-access as an example system and high throughput as an example performance metric. Such a scheme allows the decision of resource management to be made in a central processor and base stations in different timescales and allows the knowledge to be predicted with less training samples. Then, we design a deep neural network to learn the future knowledge required for making decision directly from different types of past data with different resolutions observable in cellular networks. Simulation results show that the proposed scheme with the end-to-end knowledge prediction performs closely to the relevant optimal solution with perfect and fine-grained prediction, and provides dramatic gain over non-predictive counterpart in supporting high request arrival rate for the non-realtime service.