In this paper, we propose a novel framework for designing a fast convergent multi-agent reinforcement learning (MARL)-based medium access control (MAC) protocol operating in a single cell scenario. The user equipments (UEs) are cast as learning agents that need to learn a proper signaling policy to coordinate the transmission of protocol data units (PDUs) to the base station (BS) over shared radio resources. In many MARL tasks, the conventional centralized training with decentralized execution (CTDE) is adopted, where each agent receives the same global extrinsic reward from the environment. However, this approach involves a long training time. To overcome this drawback, we adopt the concept of learning a per-agent intrinsic reward, in which each agent learns a different intrinsic reward signal based solely on its individual behavior. Moreover, in order to provide an intrinsic reward function that takes into account the long-term training history, we represent it as a long short-term memory (LSTM) network. As a result, each agent updates its policy network considering both the extrinsic reward, which characterizes the cooperative task, and the intrinsic reward that reflects local dynamics. The proposed learning framework yields a faster convergence and higher transmission performance compared to the baselines. Simulation results show that the proposed learning solution yields 75% improvement in convergence speed compared to the most performing baseline.

In view of the challenges foreseen in futuristic massive IoT (mIoT) scenarios, characterized by a huge deployment of energy-constrained IoT devices, we propose an efficient next-generation multiple access (NGMA) scheme where several innovative solutions combine and integrate. First, we aim to maximize the spectral efficiency of the physical uplink-shared channel (PUSCH) by adopting the most appropriate nonorthogonal multiple access (NOMA) technique and by introducing an efficient contention-based approach for data transmission in the unused PUSCH resources. Second, on the basis of the traffic load, the proposed scheme dynamically sets the UpLink (UL) radio resources dimensioning by searching the optimal tradeoff between lowering the collision probability in the physical random access channel (PRACH) and providing enough radio resources in the PUSCH to permit the data transmission to all the succeeded access requests. Third, it is enriched by a series of strategic procedures to make feasible the traffic load estimation, even when the PRACH is in overload condition. Fourth, we exploit artificial intelligence technologies to further optimize the performances obtained. The proposed scheme is compared with other benchmark schemes available in the literature. It outperforms them in the most important metrics for the mIoT scenario (i.e., spectral efficiency and energy consumption), especially in high load conditions, which is the expected requirement for Beyond 5G and 6G networks.

To tackle the heterogeneous requirements of beyond 5G (B5G) and future 6G wireless networks, conventional medium access control (MAC) procedures need to evolve to enable base stations (BSs) and user equipments (UEs) to automatically learn innovative MAC protocols catering to extremely diverse services. This topic has received significant attention, and several reinforcement learning (RL) algorithms, in which BSs and UEs are cast as agents, are available with the aim of learning a communication policy based on agents' local observations. However, current approaches are typically overfitted to the environment they are trained in, and lack robustness against unseen conditions, failing to generalize in different environments. To overcome this problem, in this work, instead of learning a policy in the high dimensional and redundant observation space, we leverage the concept of observation abstraction (OA) rooted in extracting useful information from the environment. This in turn allows learning communication protocols that are more robust and with much better generalization capabilities than current baselines. To learn the abstracted information from observations, we propose an architecture based on autoencoder (AE) and imbue it into a multi-agent proximal policy optimization (MAPPO) framework. Simulation results corroborate the effectiveness of leveraging abstraction when learning protocols by generalizing across environments, in terms of number of UEs, number of data packets to transmit, and channel conditions.

With the emerging of the massive Machine Type Communication usage scenario, the legacy Random Access (RA) procedures need to be fully renewed to meet the huge number of accesses and the energy-constrained device requirements. In this context, the strategy of transmitting tagged preambles as access requests is gaining importance. In fact, it can reduce the number of signaling transmissions per access attempt, the overall number of phases in the access procedure, and, accordingly, energy consumption per device. However, the effectiveness of detection strategies of tagged preambles by the gNB receiver plays a fundamental role on the RA procedure performance. Up to now, the performance of these advanced detection procedures has been obtained only by running a large number of simulations that are typically highly time-consuming. To the best of our knowledge, this paper presents the first analytical model to analyze the correct detection of both the preamble and the tag transmitted by each device in the presence of interference, due to other preambles and tags, and of noise. The high accuracy of the proposed model is verified through simulations. In addition, we show how our analytical study can be a good tool to investigate and derive innovative detection strategies.

The massive machine-type communication (mMTC) usage scenario, also known as the massive Internet of Things (mIoT), involves a huge number of MTC devices having high requirements on increased battery lifetime, which autonomously transfer small amounts of data with relaxed delay requirements, without human intervention. The current cellular network is unsuitable for this scenario, due to the limited uplink resources allocated to the physical random access channel (PRACH) and to the physical uplink shared channel (PUSCH). With this in mind, in this article, we propose a new framework, customized for massive MTC services, that includes a joint control of the dynamic resource allocation between the PRACH and the PUSCH, and a new random access (RA) procedure based on an adaptive access class barring (ACB) scheme that appropriately spreads RA reattempts in time. In addition, to further increase the transmission efficiency, we adopt the sparse code multiple access (SCMA) technique for PUSCH resources, because SCMA results as the most promising nonorthogonal multiple access (NOMA) technique to support massive MTC connectivity with small-size data. The simulation results show that the proposed control framework significantly improves the number of succeeded communications in comparison with the other proposals available in the literature. Furthermore, an energy consumption model is introduced to show the considerable energy saving achieved by the MTC devices which adopt our joint control scheme.

In the next few years, a considerable innovation concerning the design of the future 5G mobile networks will be a concrete step towards enabling effective high throughput and low latency services. Software Defined Networking (SDN), Network Function Virtualization (NFV) and Self Organizing Network (SON) are considered the enabling technologies to achieve these goals. In this paper, assuming a Control-Data Separation Architecture (CDSA), we propose a flexible SDN/NFV-based SON testbed, for future 5G mobile networks. The main contribution of our work is to cover the need for a CDSA based testbed, enabling the investigation of the NG-SON capabilities for practical implementations. We implement two different testbed setups, a real one and a virtualized one, both based on the FlexRAN and OpenAir-Interface software tools. First, we implement a specific case study, i.e., the RAN entities activation/deactivation procedures. Next, we carry out time measurements, concerning the aforementioned procedures, in order to prove proper testbed functioning. Finally, we validate the C-SON and D-SON capabilities of our testbed, considering the features of the results.

The Phantom cell concept is a solution for high-traffic outdoor environments that can also support good mobility and connectivity in ultra-dense network. In a phantom cells network, a centralized Macrocell controls all phantom cells within its coverage area and the C-Plane and U-Plane are split. Many issues need to be addressed for this architecture including radio resources allocation and cross/co tier interference mitigation. One of effective techniques used to mitigate interference is to group the cells in community (clusters). In this line, we propose a review of the max-min distance algorithm to establish the optimal cluster number, which affects significantly the system performance. Also, to assign the spectral resources to cells of the cluster, a new allocation algorithm based on traffic demand is proposed.

With the increment in users' data demand and the emergence of applications with stringent Quality of Service requirements, important improvements in cellular network architecture need to be made. In this paper we explore the symbiosis of three key communication technologies in 5G: Device-to-Device communications, 60 GHz unlicensed band transmissions and adaptive beamforming techniques. It provides notable energy saving, improvement in the system spectral efficiency, high achievable data rate and strong interference reduction. Despite these benefits, there are some issues to be analyzed: a limited communication range due to the high path loss, and the interferences among D2D communications, especially if users' density is high. To manage these problems, on the basis of a mmWave network architecture for high TIE density indoor environments, we propose a new centralized control for radio access and an efficient multi-criteria scheduling algorithm using greedy graph vertex-coloring techniques. We aim to enhance transmission efficiency, exploiting concurrent transmissions, maximize throughput and minimize end-to-end delay, while taking into account to keep the computational load low.

The rapid evolution of mobile computing, together with the spread of social networks is increasingly moving the role of users from simple information and services consumers to actual producers. Currently, while most of the critical aspects related to User-Generated Contents (UGC) have been addressed, many issues related to service generation still must be faced and represent the next challenge. In this work, we focus on security issues raised by a particular kind of services: those generated by users. User-Generated Services (UGS) are characterized by a set of features that distinguish them from conventional services. To cope with UGS security problems we introduce three possible policy management models, analyzing benefits and drawbacks of each approach. Finally, we propose a cloud-based solution that enables the composition of multiple UGS and policy models, allowing user's devices to share features and services among them.

Mobile users in vehicular environments such as trains, buses and other public transportation vehicles suffer from the low quality signal from the outside wireless network. The rapid growth in the deployment of LTE femtocells for indoor use and their benefits have led many authors to propose using them even in vehicles, implementing the so-called Moving Networks. This paper shows that the use of pure LTE femtocells exhibits relevant issues in terms of interference and consequently performance in a realistic use. In order to overcome these issues, we propose to adopt the millimeter Wave technology in the Moving Networks, creating the Hybrid Mobile Femtocells. In the paper we discuss the concerns arising from applying mmWave communications at 60 GHz inside vehicles and we provide a new throughput analysis in order to benchmark our proposed architecture to the existing implementations.

In this paper, we propose a self-optimized coverage function for LTE femtocells embedded in a macrocell area. Each Femto Base Station (FBS) adapts its pilot power, and thus the coverage, to the on-site traffic demand. Under low traffic conditions the FBSs, whose presence is not essential for the proper operation of the network, reside in a low power Listen Mode. In this way a relevant energy saving on entire femtocell network can be achieved. In a high-load scenario, FBSs dynamically create high capacity zones under interference constraints. This permits to improve system capacity and offload more traffic from the nearby macrocell and, in the same time, to minimize co-channel interference in the femtocell tier.

This paper deals with a novel buffer management scheme based on evolutionary computing for shared-memory asynchronous transfer mode (ATM) switches. The philosophy behind it is adaptation of the threshold for each logical output queue to the real traffic conditions by means of a system of fuzzy inferences. The optimal fuzzy system is achieved using a systematic methodology, based on genetic algorithms (GAs), which allows the fuzzy system parameters to be derived for each switch size, offering a high degree of scalability to the fuzzy control system. Its performance is comparable to that of the push-out (PO) mechanism, which can be considered ideal from a performance viewpoint, and at any rate much better than that of threshold schemes based on conventional logic. In addition, the fuzzy threshold (FT) scheme is simple and cost-effective when implemented using VLSI technology

This paper deals with a novel buffer management scheme based on the combination of evolutionary computing and fuzzy logic for shared-memory packet switches. The philosophy behind it is adaptation of the threshold for each logical output queue to the real traffic conditions by means of a system of fuzzy inferences. The optimal fuzzy system is achieved using a systematic methodology based on Genetic Algorithms for membership-function selecting and tuning. This methodology approach allows the fuzzy system parameters to be automatically derived when the switch parameters vary, offering a high degree of scalability to the fuzzy control system. Its performance is close to that of the push-out mechanism, which can be considered ideal from a performance viewpoint, and at any rate much better than that of threshold schemes based on conventional logic. In addition, the fuzzy threshold scheme is simple to implement, unlike the push-out mechanism which is not practically feasible in high-speed switches due to the amount of time required for computation, and above all inexpensive when implemented using current standard technology.

This paper deals with a novel buffer management scheme based on evolutionary computing for shared-memory ATM switches. The philosophy behind it is adaptation of the threshold for each output logical queue to the real traffic conditions by means of a system of fuzzy inferences. The optimal fuzzy system is achieved using a systematic methodology based on genetic algorithms for membership-function selecting and tuning. This methodology approach allows the fuzzy system parameters to be automatically derived when the switch parameters vary, offering a high degree of scalability to the fuzzy control system. Its performance is very close to that of an ideal mechanism like the push-out mechanism, and at any rate much better than that of the threshold schemes based on conventional logic. In addition it is simple to implement and above all inexpensive when implemented using VLSI technology.

We address the relevant issue of managing traffic flows with different priorities in packet switched networks, namely ATM and IP networks. We consider a reference model in which two traffic flows with different priorities are multiplexed within the buffer of a cell-based switch. The solution we propose, based on the fuzzy system theory, is able to guarantee the QoS requirements of high-priority traffic flow, allowing at the same time the exploitation of unused buffer resources to accommodate low-priority traffic flow in order to maximize the total throughput. The performance assessment of the fuzzy scheme demonstrates that our solution outperforms other popular mechanisms, based on conventional logic, such as threshold and push out mechanisms.

We assess the application of artificial intelligence techniques to the complex problem of traffic control in ATM networks. The paper deals with the close link between call admission control and usage parameter control and proposes a simulation-based analysis to demonstrate how inefficiency on the part of policing affects bandwidth allocation. To take this into account, the paper proposes a framework for traffic control in which the CAC and policing functions are both based on artificial intelligence techniques, i.e. neural networks and fuzzy logic. In this way it is possible to train the neural network in such a way as to take into account the real behavior of the policer. As the results obtained show this allows us to implement traffic management strategies which can improve the exploitation of network resources.

Concerns a fuzzy logic-based system which has been purposely designed to achieve real-time traffic control in high-speed networks using the asynchronous transfer mode (ATM) technique. One of the most critical functions is "policing", which has the task of ensuring that each user source complies with the traffic parameters negotiated in the call setup to avoid network congestion. This function is difficult to implement on account of certain conflicting requirements such as selectivity and responsiveness. This is confirmed by the severe limits affecting the most popular mechanisms proposed so far, based on conventional logic. The capacity to formalize approximate reasoning processes offered by fuzzy logic is exploited to derive rules of behavior for a policer starting from the know-how of an expert. We address two key issues related to the implementation of the fuzzy policer. The first focuses on the possibility of hardware implementation of the mechanism using VLSI technology; we present the design of a VLSI fuzzy processor which exhibits a level of performance of over 3 MFLIPS. The second issue concerns the suitability of applying the fuzzy policer to the policing of several classes of sources to reach high levels of cost effectiveness and scalability.

In ATM networks, usage parameter control is required in order to ensure that each source conforms to its negotiated parameters. To this purpose, several policing methods, such as leaky bucket and window mechanisms, have been introduced in literature. However, traditional methods have proved to be inefficient in coping with the conflicting requirements of ideal policing, that is, a low false alarm probability and high responsiveness. This led us to explore alternative solutions based on artificial intelligence techniques, specifically, in the field of fuzzy systems. We propose a policing mechanism based on fuzzy logic that aims at detecting violations of the parameters negotiated. The main characteristics of the proposed fuzzy policer are simplicity and the capacity to combine a high degree of responsiveness with a selectivity close to that of an ideal policer. Moreover, it can easily be implemented in hardware, thus, enhancing both cost and processing performance. The reported simulation results show that the performance of our fuzzy policer is much better than that of conventional policing mechanisms.

Most of the policing techniques proposed so far in high speed networks using the ATM technique are based on conventional approaches which use a crisp decision making logic. They don't meet the selectivity and responsiveness requirements needed to make a policing efficient. In this paper we propose a policing mechanism based on fuzzy logic. The main characteristics of the proposed fuzzy policer are simplicity and the capacity to combine a high degree of responsiveness with a selectivity close to that of an ideal policer. Moreover, it can easily be implemented in hardware, thus enhancing both cost and processing performance. The simulation results reported show that the performance of our fuzzy policer is much better than that of conventional policing mechanisms.

Traffic management in ATM networks aims for the controlled use of network resources to prevent the network from becoming a bottleneck. This is a very complex task due to the random nature of the traffic arrival pattern and contention for network resources. In this paper we deal with one of the most important aspects of traffic control: "the policing". It has the role of ensuring that each source complies with the traffic parameter negotiated, in order to avoid network congestion. We present a policing mechanism based on Fuzzy Logic. As far as selectivity and dynamic response are concerned, it outperforms traditional mechanism thanks to the softness which characterizes Fuzzy Logic. In the paper we also address the design of a VLSI fuzzy processor to implement the mechanism in hardware. The processor is organized as a cascade of pipeline stages in which fuzzy inferences are processed in parallel. A synthesis on CMOS 0.5 /spl mu/m technology has allowed the processor to reach a very high processing speed occupying only a few mm/sup 2/ of silicon area, thus requiring a very low implementation cost.

Erasure nodes are special devices that have been conceived to improve the system throughput of distributed queue dual bus (DQDB) networks. The authors discuss the overload performance of a network with erasure nodes and the standard bandwidth balancing (BWB) mechanism enabled, when alternative implementation schemes of the erasure node protocol are considered. A numerical method is presented to calculate the maximum throughput in any node of the network, and the conditions for achieving fairness are discussed. The results obtained through the analytical approach are assessed by simulation.

The fair queue dual bus (FQDB) medium-access control (MAC) protocol for dual bus networks, which is intrinsically fair in bandwidth sharing and retains all the positive features of distributed-queue dual-bus (DQDB) systems including simplicity, robustness, flexibility, 100% bandwidth utilization, and zero access delay at very low load, is presented. Features which make this protocol appealing with respect to others DQDB alternatives are represented by the fact that FQDB is compatible with the DQDB standard with respect to the physical arrangement and attachment to the bus, the XOR write tap, the access control field (ACF) length, and the three priority levels. Nevertheless, its properties are attractive also for protocols operating at 1 Gb/s and up. Fairness is formally proved, and it is shown that a bandwidth sharing pattern different from the uniform one can easily be enforced, as may be the case if bridges or other large users are presented. Queueing disciplines, priorities, and implementation issues are discussed, and comparisons obtained by simulation are given.>

The authors deal with a new technique that is fully compatible with the IEEE 802.6 MAN standard and allows for multiple use of a slot for segment transmissions. In particular, this technique increases the efficiency of the distributed-queue dual-bus (DQDB) protocol by allowing pre-use of slots traveling along the bus before they are taken as determined by normal DQDB operation. These slots may be captured by nodes wishing to transmit their queued-arbitrated segments addressed to neighbors. The proposed technique is particularly well suited for the case when information transfer between neighboring nodes is more likely than that between remote ones. Simulation was used to study the throughput, delay, and fairness performance. It is demonstrated that the resulting extension of DQDB, which is called SP-DQDB (slot pre-use DQDB), may be much more efficient than DQDB in several different environments.>

The authors introduce a control strategy for bandwidth management in asynchronous transfer mode (ATM) networks. A two-level hierarchy is defined, where one level performs fixed class-selective call admission control strategies that are periodically dynamically coordinated by a higher-level bandwidth allocation controller. Each admission controller decides to accept or refuse an incoming call on the basis of a class-selective rule designed to maintain a certain quality of service. A decision is taken according to the virtual capacity share that is assigned to the various service classes by the allocation controller. Bandwidth shares are periodically recomputed online by means of the constrained minimization of a cost function that takes cell loss probability and refused traffic into account. The control structure, the strategies, and the optimization algorithm used are described as well as the assumptions underlying the choices made. Initial simulation results are reported.>

The authors present the delay-throughput performance of a new MAC protocol, the fair queue dual bus (FQDB), which relieves dual bus networks from the unfairness problems exhibited by the basic version of the distributed queue dual bus (DQDB) standard. The protocol considered is absolutely fair in bandwidth sharing and does not imply any change in the DQDB transmission system. It retains all the positive features of DQDB including simplicity, robustness, flexibility, 100% bandwidth utilization, and little slot access delay at light load. Comparisons to DQDB with the bandwidth balancing mechanism show that the FQDB delay-throughput performance was always superior while its implementation complexity was comparable.>

An architecture and a media access control (MAC) protocol aimed at improving fairness in dual-bus networks are introduced. The major feature of this proposal is that the distributed operation of updating the queue is made atomic, so that the average bus access delay is equalized over the network. An equalized access gain is provided by mandating that a station waiting for transmission consider itself as positioned in the distributed queue only after a constant delay, which is needed for all the other stations to register the request of the first one. This delay is equal to the propagation time of the request along the entire bus. It is shown that under certain conditions the average access delay results in the one experienced in the distributed queue dual bus (DQDB) by those stations which are placed at the middle of the bus. Consequently, the global end-to-end delay remains much below the limits fixed by the CCITT for delay sensitive services. The actual average delay is even smaller than the one estimated due to the presence in the protocol of the stand-by mechanism which is the same as in the DQDB version. In comparison with the traditional DQDB, the increase in access-delay values is rewarded by the significant improvement in terms of fairness.>