Next-generation wireless networks are expected to support extremely high data rates and radically new applications, which require a new wireless radio technology paradigm. The challenge is that of assisting the radio in intelligent adaptive learning and decision making, so that the diverse requirements of next-generation wireless networks can be satisfied. Machine learning is one of the most promising artificial intelligence tools, conceived to support smart radio terminals. Future smart 5G mobile terminals are expected to autonomously access the most meritorious spectral bands with the aid of sophisticated spectral efficiency learning and inference, in order to control the transmission power, while relying on energy efficiency learning/inference and simultaneously adjusting the transmission protocols with the aid of quality of service learning/inference. Hence we briefly review the rudimentary concepts of machine learning and propose their employment in the compelling applications of 5G networks, including cognitive radios, massive MIMOs, femto/small cells, heterogeneous networks, smart grid, energy harvesting, device-todevice communications, and so on. Our goal is to assist the readers in refining the motivation, problem formulation, and methodology of powerful machine learning algorithms in the context of future networks in order to tap into hitherto unexplored applications and services.

/pdf/machine-learning-paradigms-for-next-generation-wireless-pk4ulk0xuk.pdf

Machine Learning Paradigms for Next-Generation Wireless Networks

This paper provides an overview of cognitive radio (CR) networks, with focus on the recent advances in resource allocation techniques and the CR networks architectural design. The contribution of this work is threefold. First, a systematic way to study the resource allocation problem is presented; various design approaches are introduced, such as signal-to-interference-and-noise ratio (SINR) or transmission power-based, and centralized or distributed methods. Second, CR optimization methods are presented, accompanied by a comprehensive study of the resource allocation problem formulations. Furthermore, quality of service criteria of the physical or/and the medium access control layers are investigated. Third, challenges in spectrum assignment are discussed, focusing on dynamic spectrum allocation, spectrum aggregation and frequency mobility. Such approaches constitute an emerging trend in efficient spectrum sharing and affect the performance of resource allocation techniques. The open issues for future research in this area are finally discussed, including adaptability-reconfigurability, dual accessibility, and energy efficiency.

Radio Resource Allocation Techniques for Efficient Spectrum Access in Cognitive Radio Networks

While machine learning and artificial intelligence have long been applied in networking research, the bulk of such works has focused on supervised learning. Recently, there has been a rising trend of employing unsupervised machine learning using unstructured raw network data to improve network performance and provide services, such as traffic engineering, anomaly detection, Internet traffic classification, and quality of service optimization. The growing interest in applying unsupervised learning techniques in networking stems from their great success in other fields, such as computer vision, natural language processing, speech recognition, and optimal control (e.g., for developing autonomous self-driving cars). In addition, unsupervised learning can unconstrain us from the need for labeled data and manual handcrafted feature engineering, thereby facilitating flexible, general, and automated methods of machine learning. The focus of this survey paper is to provide an overview of applications of unsupervised learning in the domain of networking. We provide a comprehensive survey highlighting recent advancements in unsupervised learning techniques, and describe their applications in various learning tasks, in the context of networking. We also provide a discussion on future directions and open research issues, while identifying potential pitfalls. While a few survey papers focusing on applications of machine learning in networking have previously been published, a survey of similar scope and breadth is missing in the literature. Through this timely review, we aim to advance the current state of knowledge, by carefully synthesizing insights from previous survey papers, while providing contemporary coverage of the recent advances and innovations.

/pdf/unsupervised-machine-learning-for-networking-techniques-1u9zi0s9j5.pdf

Unsupervised Machine Learning for Networking: Techniques, Applications and Research Challenges

Currently, the radio spectrum is statically allocated and divided between licensed and unlicensed frequencies. Due to this inflexible policy, some frequency bands are growing in scarcity, while large portions of the entire radio spectrum remain unused independently of time and location. Cognitive Radio is a recent network paradigm that aims a more flexible and efficient usage of the radio spectrum. Basically, it allows wireless devices to opportunistically access portions of the entire radio spectrum without causing any harmful interference to licensed users. The present document surveys the literature on Cognitive Radio. It aims to provide a comprehensive and self-contained description of this research topic area, mainly focusing on communication protocols, spectrum decision issues, and future research directions. It is a tutorial in nature and consequently does not require any previous knowledge about Cognitive Radio. Readers are only required to have some general background on wireless data networks. Emphasis is put on Cognitive Radio genesis, issues that must be addressed, related technologies, standardization efforts, the state of the art, and future research directions according to the vision of the authors.

Cognitive radio: survey on communication protocols, spectrum decision issues, and future research directions

The cognitive radio is an emerging technology that enables dynamic spectrum access in wireless networks. The cognitive radio is capable of opportunistically using the available portions of a licensed spectrum to improve the application performance for unlicensed users. The opportunistic use of the available channels in the wireless environment requires dynamic channel assignment to efficiently utilize the available resources while minimizing the interference in the network. A challenging aspect of such algorithms is the incorporation of the channels' diverse characteristics, highly dynamic network conditions with respect to primary users' activity, and different fragmented sizes of the available channels. This paper presents a comprehensive survey on the state-of-the-art channel assignment algorithms in cognitive radio networks. We also classify the algorithms by presenting a thematic taxonomy of the current channel assignment algorithms in cognitive radio networks. Moreover, the critical aspects of the current channel assignment algorithms in cognitive radio networks are analyzed to determine the strengths and weaknesses of such algorithms. The similarities and differences of the algorithms based on the important parameters, such as routing dependencies, channel models, assignment methods, execution model, and optimization objectives, are also investigated. We also discuss open research issues and challenges of channel assignment in the cognitive radio networks.

Channel Assignment Algorithms in Cognitive Radio Networks: Taxonomy, Open Issues, and Challenges

Neural network-based learning schemes for cognitive radio systems

Performance evaluation of artificial neural network-based learning schemes for cognitive radio systems

This paper proposes enhancements to the channel(-state) estimation phase of a cognitive radio system. Cognitive radio devices have the ability to dynamically select their operating configurations, based on environment aspects, goals, profiles, preferences etc. The proposed method aims at evaluating the various candidate configurations that a cognitive transmitter may operate in, by associating a capability e.g., achievable bit-rate, with each of these configurations. It takes into account calculations of channel capacity provided by channel-state estimation information (CSI) and the sensed environment, and at the same time increases the certainty about the configuration evaluations by considering past experience and knowledge through the use of Bayesian networks. Results from comprehensive scenarios show the impact of our method on the behaviour of cognitive radio systems, whereas potential application and future work are identified.

Enhancing Channel Estimation in Cognitive Radio Systems by means of Bayesian Networks

In the era of wireless communications, Beyond the 3rd Generation (B3G), a network operator (NO) should satisfy numerous requirements, namely, personalisation, context awareness, always best connectivity, ubiquitous service provision and seamless mobility. A NO can efficiently satisfy the requirements by relying on the different radio networks of its heterogeneous infrastructure, and potentially on other cooperating networks. In this respect, the NO should possess advanced management mechanisms for driving its users to the most appropriate networks that satisfy the requirements. The presentation of such a management system is the specific contribution of this paper. The system is called Reconfigurable Terminal Management System (RTMS). In general, it provides the means for profile modelling, the acquisition of monitoring/discovery/context information, and the negotiation and selection of configurations, based on information deriving from policies, as well as the profiles and the context. Our work focuses on the role and the information of the components of the RTMS. Concrete functionality for accomplishing the role is also presented. Nevertheless, the system is open to the integration of alternate functionality. Our discussion includes a business case that presents in high-level terms the role of the management system, a detailed description of the components of the management system and results that show the efficiency of the management schemes. A summary and further research challenges, conclude this article.

Management System for Terminals in the Wireless B3G World

A number of new, powerful, wireless networking standards have appeared during the last years. In this context, a key topic in the research area of B3G/4G networks is related to strategies and mechanisms to realise the efficient complementary use of the different technologies. The aim is to be able to make the most out of the variety of available access standards. According to the B3G concept, the diverse Radio Access Technologies (RATs) can converge into one composite radio environment, where the user is "always best connected" and can move between the various technologies seamlessly. Cognitive and reconfigurable wireless networks have emerged as a complementary concept to B3G networks. Cognitive systems determine their behaviour, in a reactive or proactive manner, based on external triggers, as well as their goals, principles (policies, rules), experience and knowledge. The focus of this paper is on a Cognitive Reconfigurable Equipment Management System (CREMS).

A. Katidiotis

Papers

Neural network-based learning schemes for cognitive radio systems

Performance evaluation of artificial neural network-based learning schemes for cognitive radio systems

Enhancing Channel Estimation in Cognitive Radio Systems by means of Bayesian Networks

Management System for Terminals in the Wireless B3G World

Evolution in equipment management concepts: from reconfigurable to cognitive wireless terminals