Search

Proactive reconfiguration of network slices according to uncertain traffic demands is essential to improve network resource utilization while ensuring service quality in 5G-and-beyond systems. Existing researches on network slice reconfiguration are either model-driven or data-driven methods. However, model-driven methods may cause resource over-provisioning due to a lack of prediction mechanism, while data-driven methods are unrealistic in inter-slice reconfiguration that involves costly and time-consuming operations such as VNF migration. To address these issues, in this paper, we propose a Hybrid Model-Data driven (HMD) framework that intelligently performs inter-slice reconfiguration by leveraging prediction interval and robust optimization. We design a Prediction Interval-oriented Predictor (PIP) to produce a prediction interval that can bracket the future traffic demand with a prespecified probability. Based on the prediction interval, we design an inter-slice reconfiguration scheme (named box optimizer) to perform fast inter-slice reconfigurations. To tackle the over-conservativeness of the box optimizer, we further design the ellipsoid optimizer with better optimality at a cost of increased complexity. Numerical results demonstrate that the proposed framework can provide high robustness with low power consumption. Meanwhile, the trade-off between the power consumption and the realized robustness can be flexibly adjusted according to the type of slice and the level of traffic demand fluctuations.

This paper studies semi-blind channel estimation for a reconfigurable intelligent surface (RIS) aided uplink massive multiple-input multiple-output (MIMO) system, in which the base station simultaneously estimates the channel coefficients and detects the partially unknown transmit symbols. We formulate the semi-blind channel estimation task as a trilinear inference problem. Based on the approximate message passing (AMP) principle, we develop a computationally efficient approach, called Trilinear AMP, to calculate the marginal posterior mean estimators of the trilinear inference problem. Simulation results demonstrate the effectiveness of the proposed Trilinear AMP approach.

To support massive access for future wireless communications, we propose a novel intelligent reflecting surface (IRS) enhanced downlink multi-user multi-input single-output (MU-MISO) symbiotic radio (SR) system, where each IRS, acting as a reflecting Internet-of-Things (IoT) device, transmits its message to a nearby primary receiver (PR) by reflecting the RF signals from the primary transmitter (PT), and simultaneously enhances the transmission from the PT to the associated PR. Thus, each PR jointly decodes its own message as well as the one from the corresponding IRS. We are interested in maximizing the weighted sum-rate of both primary and IoT transmissions by jointly designing the active transmit beamforming at PT and the passive beamforming at each IRS, subject to the maximum transmit power constraint at PT. Besides, as the passive elements at IRS can only reflect the incident signal with discrete phase shifts in practice, the discrete reflection coefficient (RC) constraint is further considered at the IRSs. Due to the non-convexity of the formulated problems, we solve them with fractional programming (FP) technique and alternating optimization (AO) method. Simulation results have verified the effectiveness of the proposed algorithms compared to different benchmark schemes.

The Open Radio Access Network (O-RAN) introducing a particular unit known as RAN Intelligent Controllers (RICs) has been regarded as revolutionary paradigms to support multiclass wireless services required in the fifth and sixth generation (5G/6G) networks. Through unprecedentedly installing various machine learning (ML) algorithms to RICs, a RAN is able to intelligently configure resources/communications to support any vertical applications over any operating scenarios. However, to practically deploy this RAN paradigm, the O-RAN still suffers two critical issues of load balance and handover control, and therefore the very first ML algorithm for the O-RAN should effectively address these issues. In this paper, inspired by the superior performance of deep reinforcement learning (DRL) in tackling sequential decision-making tasks, we therefore develop an intelligent user access control scheme with the facilitation of deep Q-networks (DQNs). A federated DRL-based scheme is further proposed to train the parameters of multiple DQNs in the O-RAN, so as to maximize the long-term throughput and meanwhile avoid frequent user handovers with a limited amount of signaling overheads in the O-RAN. The simulation results have fully demonstrated the outstanding performance over the state-of-the-arts, to service the urgent needs in the standardization of the O-RAN.

Symbiotic radio (SR) emerges as a spectrum and energy-efficient communication paradigm for future passive Internet-of-Things (IoT). In this paper, we consider a multiuser multi-backscatter-device (BD) SR communication system to enhance the spectrum efficiency, by sharing a common time and frequency resource block. However, the receiver design becomes much more complicated due to the presence of inter-user and inter-BD interference. To address these issues, we propose a novel receiver design to perform joint channel estimation, user data decoding, and BD symbol detection. Specifically, motivated by the idea of approximate message passing, we develop a computationally efficient iterative algorithm under the Bayesian inference framework to resolve the joint estimation problem. Simulation results demonstrate the effectiveness of the proposed receiver design.

In this paper, we develop a framework to optimize the trade-off between radar sensing and data transmission in Joint Radar-Communication (JRC) systems under smart and reactive jamming attacks. First, we propose a novel JRC design architecture that uses backscatter technology and deception strategy to leverage jamming attacks for a JRC system. The deception strategy is used to predict the jammer's action and adopt appropriate counterattack instantaneously, while backscatter technology is used to transmit data on the jamming signals. To deal with the jamming strategy uncertainty (e.g., jamming capability), we then develop a deep reinforcement learning algorithm to quickly find the optimal defense policy for the JRC system. Our in-depth investigation reveals that the proposed design not only significantly undermines the jamming attacks, but also utilises jamming signals to improve the system performances. Compared with conventional anti-jamming methods, our proposed design significantly improves data throughput while maintaining a satisfactory radar sensing performance in dynamic environments. Moreover, the proposed deep reinforcement learning based solution can converge four times faster than a conventional deep Q- Learning based solution.

This paper considers a reconfigurable intelligent surface (RIS) enhanced symbiotic radio (SR) multigroup multicast system, where the RIS passively transmits information to an Internet-of-Things receiver (IR) by modulating the incident multicasting signals, and also enhances the downlink multicast transmission from a base station (BS) to multiple primary receivers (PRs). We aim to minimize the BS’s transmit power by jointly optimizing the active transmit beamforming at the BS and the passive phase shifts at the RIS, subject to the constraints on the PRs’ signal-to-interference-plus-noise-ratio, the IR’s successive-interference-cancellation and signal-to-noise-ratio, as well as the RIS’s phase shifts. An efficient iterative algorithm is proposed to solve the non-convex problem, by employing alternating optimization, quadratic transform and semidefinite relaxation. The algorithm’s convergence and complexity are also analysed. Numerical results show that the proposed scheme consumes less power than the traditional multicast scheme without RIS.

In this paper, we investigate the resource allocation problem for the symbiotic radio system, in which the passive backscatter transmission is parasitic in the primary transmission, and the cooperative receiver simultaneously detects the signals transmitted from the active primary transmitter (PT) and the backscatter device (BD). In particular, channel fading is considered, and the ergodic weighted sum rate of the primary and backscatter transmissions is maximized by jointly optimizing the transmit power at the PT and the reflection coefficient at the BD under either long-term or short-term transmit-power constraint over the fading states. Two practical transmission setups are considered, where the relationship between the primary and backscatter transmissions is either commensal or parasitic. For the commensal setup, we fix the reflection coefficient as the maximum value and derive the optimal power allocation scheme based on the standard convex optimization procedure. For the parasitic setup, the weighted sum-rate maximization problem is non-convex, and we resort to the concave–convex procedure to derive a suboptimal solution. Simulation results demonstrate that in the commensal setup, the BD may realize transmission while causing no harmful interference to the primary system; in the parasitic setup, a higher backscatter data rate can be achieved through slightly sacrificing the primary transmission rate.

Non-orthogonal multiple access (NOMA) is envisioned as a key technology to enhance the spectrum efficiency for 5G cellular networks. Meanwhile, ambient backscatter communication is a promising solution to the Internet of Things (IoT), due to its high spectrum efficiency and power efficiency. In this paper, we are interested in a symbiotic system of cellular and IoT networks and propose a backscatter-NOMA system, which incorporates a downlink NOMA system with a backscatter device (BD). In the proposed system, the base station (BS) transmits information to two cellular users according to the NOMA protocol, while a BD transmits its information over the BS signals to one cellular user using the passive radio technology. In particular, if the BS only serves the cellular user that decodes BD information, the backscatter-NOMA system turns into a symbiotic radio (SR) system. We derive the expressions of the outage probabilities and the ergodic rates and analyze the corresponding diversity orders and slopes for both backscatter-NOMA and SR systems. Finally, we provide the numerical results to verify the theoretical analysis and demonstrate the interrelationship between the cellular networks and the IoT networks.

The accurate identification of congested links is crucial for network performance monitoring. Network boolean tomography uses end-to-end path measurements to identify congested links, and appears as a significant alternative when direct link monitoring is not available. However, most of existing tomographic methods assume no correlations between links, i.e., the congestion of one link is assumed to be independent from the congestion of any others, hindering their applications in practice because links could become correlated during a joint optimization procedure of many network operations like traffic routing and balancing. In this paper, we study practical network boolean tomography without such an assumption. We elaborate on the ill-posed nature of network boolean tomography to highlight the significance of integrating link correlations, and model the congested link identification from end-to-end congestion observations of paths as a problem of Maximum A-Posteriori (MAP) estimation. To avoid the explicit acquisition of any priori knowledge of link correlations, we then propose a learning-based algorithm with Long Short-term Memory (LSTM), a special recurrent neural network that is good at learning statistical dependencies of sequence elements from historical data. Numerical results over real network topologies validate our learning-based network boolean tomography.

In this paper, we investigate a large intelligent surface (LIS) assisted symbiotic radio (SR) system, in which a LIS device, operating as an Internet-of-Things (IoT) device, exploits the signal from a primary transmitter (PT) as its communication carrier to achieve its own information transmission, and concurrently serves as a desirable additional link to aid the primary transmission from the PT to a primary receiver (PR). A cooperative beamforming scheme (i.e., active transmit beamforming at the PT and passive reflecting beamforming at the LIS device) is proposed to minimize PT's transmit power under quality-of-service (QoS) constraints of both the primary and LIS device transmissions. Both continuous and discrete phase shift setups of the LIS device are considered. For the continuous phase shift setup, a closed-form solution is derived, analytically showing that by smartly configuring the phase shifts, the signals from primary link and backscatter link can add coherently at the PR; while for the discrete phase shift setup, a near-optimal solution for the 1-bit phase shifter is obtained via the semi-definite relaxation (SDR) technique, and a general successive refinement algorithm (SRA) is developed for any-bit phase shifter. Simulation results demonstrate that cooperative beamforming design can adaptively adjust beamformers to strike a balance between the primary and LIS device transmissions.

Reconfigurable Intelligent Surface (RIS) is a new and promising technique to solve the energy-efficiency, spectral-efficiency and hardware-cost problem faced by beyond-5G wireless networks. In this paper, we consider a RIS-assisted two-way relay network in which two users exchange information via the base station (BS) with the help of a RIS. By jointly designing the beamforming matrix at the BS and the phase shifts introduced by the RIS, the minimum SNR of the two users is maximized, under the transmit power constraint at the BS. The formulated problem is non-convex and difficult to solve in general. To start with, we first study the single BS antenna case. The design problem is reduced to the problem of how to choose the phase shifts for the RIS, in addition to optimizing the BS power amplification parameter. A Channel-Gain-Maximization (CGM) algorithm is proposed to solve the problem. For the multiple BS antenna case, we decouple the phase shifts and the beamforming matrix by taking an upper bound of the SNR. The way to obtain the RIS phase shifts is similar to the single BS antenna case while the beamforming matrix is obtained by utilizing an existing solution. A Channel-Gain-Maximization Maximal-Ratio-Beamforming (CGM-MRB) algorithm is thus developed. Finally, numerical results are presented to show the effectiveness of the proposed algorithms.

Network slicing is identified as a fundamental architectural technology for future mobile networks since it can logically separate networks into multiple slices and provide tailored quality of service (QoS). However, the introduction of network slicing into radio access networks (RAN) can greatly increase user handover complexity in cellular networks. Specifically, both physical resource constraints on base stations (BSs) and logical connection constraints on network slices (NSs) should be considered when making a handover decision. Moreover, various service types call for an intelligent handover scheme to guarantee the diversified QoS requirements. As such, in this article, a multi-agent reinforcement LEarning based Smart handover Scheme, named LESS, is proposed, with the purpose of minimizing handover cost while maintaining user QoS. Due to the large action space introduced by multiple users and the data sparsity caused by user mobility, conventional reinforcement learning algorithms cannot be applied directly. To solve these difficulties, LESS exploits the unique characteristics of slicing in designing two algorithms: 1) LESS-DL, a distributed ${Q}$ -learning algorithm to make handover decisions with reduced action space but without compromising handover performance; 2) LESS-QVU, a modified ${Q}$ -value update algorithm which exploits slice traffic similarity to improve the accuracy of ${Q}$ -value evaluation with limited data. Thus, LESS uses LESS-DL to choose the target BS and NS when a handover occurs, while ${Q}$ -values are updated by using LESS-QVU. The convergence of LESS is theoretically proved in this article. Simulation results show that LESS can significantly improve network performance. In more detail, the number of handovers, handover cost and outage probability are reduced by around 50%, 65%, and 45%, respectively, when compared with traditional methods.

Existing tag signal detection algorithms inevitably suffer from a high bit error rate (BER) due to the difficulties in estimating the channel state information (CSI). To eliminate the requirement of channel estimation and to improve the system performance, in this paper, we adopt a deep transfer learning (DTL) approach to implicitly extract the features of communication channel and directly recover tag symbols. Inspired by the powerful capability of convolutional neural networks (CNN) in exploring the features of data in a matrix form, we design a novel covariance matrix aware neural network (CMNet)-based detection scheme to facilitate DTL for tag signal detection, which consists of offline learning, transfer learning, and online detection. Specifically, a CMNet-based likelihood ratio test (CMNet-LRT) is derived based on the minimum error probability (MEP) criterion. Taking advantage of the outstanding performance of DTL in transferring knowledge with only a few training data, the proposed scheme can adaptively fine-tune the detector for different channel environments to further improve the detection performance. Finally, extensive simulation results demonstrate that the BER performance of the proposed method is comparable to that of the optimal detection method with perfect CSI., Accepted by IEEE Globecom 2020; Journal version "Deep Transfer Learning for Signal Detection in Ambient Backscatter Communications" has been accepted by IEEE TWC. arXiv admin note: substantial text overlap with arXiv:2009.05231

It is widely acknowledged that network slicing can tackle the diverse usage scenarios and connectivity services that the 5G-and-beyond system needs to support. To guarantee performance isolation while maximizing network resource utilization under dynamic traffic load, network slice needs to be reconfigured adaptively. However, it is commonly believed that the fine-grained resource reconfiguration problem is intractable due to the extremely high computational complexity caused by numerous variables. In this paper, we investigate the reconfiguration within a core network slice with aim of minimizing long-term resource consumption by exploiting Deep Reinforcement Learning (DRL). This problem is also intractable by using conventional Deep Q Network (DQN), as it has a multi-dimensional discrete action space which is difficult to explore efficiently. To address the curse of dimensionality, we propose a discrete Branching Dueling Q-network (discrete BDQ) by incorporating the action branching architecture into DQN, for drastically decreasing the number of estimated actions. Based on the discrete BDQ network, we develop an intelligent network slice reconfiguration algorithm (INSRA). Extensive simulation experiments are conducted to evaluate the performance of INSRA and the numerical results reveal that INSRA can minimize the long-term resource consumption and achieve high resource efficiency compared with several benchmark algorithms.

To manage the interference and enhance the sum-rate of the heterogeneous network (HetNet), conventional power control algorithms first collect instantaneous global channel state information (CSI) and then design sub-optimal power control solutions. But, the global CSI in the HetNet typically changes fast and it is demanding to collect instantaneous global CSI. In this paper, we exploit deep reinforcement learning to design a multi-agent non-cooperative power control algorithm. Particularly, a deep neural networks (DNN) is established at each access point (AP) and a multiple-actor-shared-critic (MASC) method is developed to effectively train the DNNs. Then, each AP can use the DNN to independently optimize the transmit power by feeding only local information into the DNN. Simulation results show that, the proposed algorithm can rapidly converge to an average sum-rate higher than those of conventional power control algorithms.

In this paper, we study the cross-layer performance of a symbiotic system comprising passive internet-of-things (IoT) devices and the ambient system using carrier sense multiple access (CSMA) with collision notification (CSMA/CN) MAC protocol. Different from the CSMA with collision avoidance (CSMA/CA), the CSMA/CN protocol can enhance the throughput of the ambient system notably by using physical layer techniques to detect collisions. We first evaluate the physical-layer outage probabilities for the ambient and the backscatter device (BD) system respectively, and then introduce the outage probability of the ambient system into the MAC-layer performance analysis. Thereafter, the crosslayer outage capacities of the ambient system and the BD system are derived by considering the available transmission time for the BD system. We find that the overall system performance can be improved by appropriately setting the reflection coefficient of the BD and the parameters of the MAC protocol. The simulation results are provided to verify our theoretical analysis and demonstrate the system performance.

Symbiotic radio (SR) is a promising technology for energy- and spectrum-efficient wireless communication, which exploits passive communication for Internet-of-Things (IoT) transmission and achieves a mutualistic spectrum sharing between the passive and active transmissions. In this paper, we study an reconfigurable intelligent surface (RIS) empowered symbiotic radio over a broadcasting system, i.e., a base station (BS) broadcasts signals to multiple primary receivers (PRs) under the assistance of an RIS, while the RIS also transmits information to an IoT receiver (IR) by riding over the broadcasting signals. We formulate a problem to minimize the BS’s transmit power by jointly optimizing the BS’s active precoding and the RIS’s passive beamforming, under the signal-to-noise-ratio constraints of the primary and IoT transmissions. However, the problem is challenging to be solved optimally, since the variables are coupled and the constraints are non-convex. An iterative algorithm based on block coordinated descent (BCD) and semidefinite relaxation (SDR) techniques is first proposed, and its convergence together with complexity are analyzed. Then, to tackle the problem of high computational complexity caused by SDR technique, we further propose an alternative algorithm based on generalized power method (GPM) technique. Simulation results validate that the proposed system outperforms the traditional broadcasting system without RIS. The GPM-based algorithm achieves nearly the same transmit power performance as SDR-based algorithm, with a significantly reduced computational complexity.

Conventionally, a substantial number of reflecting elements (REs) is deployed at the intelligent reflecting surface (IRS) to mitigate the effect of the double-fading attenuation in the IRS-aided link, leading to a large surface size and considerable power consumption. In this paper, a new type of IRS, called active IRS, is proposed to solve this challenge by allowing each RE to amplify the incident signal with the assistance of the active loads (negative resistances). Thus, given a power budget at the IRS, the IRS-aided link can be enhanced by increasing the number of active REs as well as amplifying the incident signal. Specifically, we consider the use of an active IRS-aided single input multiple output (SIMO) system, in which the received signal-to-noise ratio (SNR) is maximized, by optimizing not only the reflecting coefficient matrix at the IRS but also the receive beamforming at the receiver. To solve this non-convex problem, we propose an alternating optimization algorithm, that iteratively optimizes the two design variables. In particular, the receive beamforming is founded to be in the form of a linear minimum mean square error (MMSE) detector, and the reflecting coefficient matrix is obtained via the Charnes-Cooper transformation and the semi-definite programming (SDP). Simulation results show that under a practical power consumption model, the proposed active IRS-aided system achieves better performance over the conventional passive IRS-aided system with the same power budget.

Due to the explosive growth of data traffic, and poor indoor coverage, ultra-dense network (UDN) has been introduced as a fundamental architectural technology for 5G-and-beyond systems. As the telecom operator is shifting to a plug-and-play paradigm in mobile networks, network planning, and optimization become difficult, and costly, especially in residential small-cell base stations (SBSs) deployment. Under this circumstance, severe inter-cell interference (ICI) becomes inevitable. Therefore, interference mitigation is of vital importance for indoor coverage in mobile communication systems. In this paper, we propose a fully distributed self-learning interference mitigation (SLIM) scheme for autonomous networks under a model-free multi-agent reinforcement learning (MARL) framework. In SLIM, individual SBSs autonomously perceive surrounding interferences, and determine downlink transmit power without necessity of signaling interactions between SBSs for mitigating interferences. To tackle the dimensional disaster of joint action in the MARL model, we employ the Mean Field Theory to approximate the action value function to greatly decrease the computational complexity. Simulation results based on 3GPP dual-stripe urban model demonstrate that SLIM outperforms several existing known interference coordination schemes in mitigating interference, and reducing power consumption while guaranteeing UEs’ quality of service for autonomous UDNs.