Elsevier

Physical Communication

Volume 43, December 2020, 101190
Physical Communication

Full length article
Enhanced cooperative behavior and fair spectrum allocation for intelligent IoT devices in cognitive radio networks

https://doi.org/10.1016/j.phycom.2020.101190Get rights and content

Abstract

The extensive utilization of Internet of Things (IoT) devices in a wide range of services prompts spectrum scarcity. Recently, Cognitive Radio Networks (CRNs) have shown promising results in addressing the spectrum scarcity problem. Therefore, CRNs are envisaged to play a crucial role in accommodating the massive number of emerging IoT devices. However, two dominant shortcomings hinder the coalescence of IoT and CRNs. First, the selfish behavior among Secondary Users (SUs) during the spectrum sensing process. Second, the inability to construct a convenient metric for fair spectrum assignment. The contribution of this paper is quadruple: (i) a Direct Strategy Selection (DSS) mechanism that satisfies Aspiration Equilibrium (AE) and enforces cooperation among SUs during the spectrum sensing process; (ii) a spectrum allocation scheme that takes cooperative behavior as a fairness parameter; (iii) studying the effects of three windowing techniques on the proposed allocation scheme and selecting the window that achieves maximum fairness; (iv) exploring the trade-off between the utility function and cooperation among SUs. The results show that DSS yields instant convergence and a threefold increase in the index of cooperation. Regarding spectrum allocation, the proposed scheme combined with the Reverse Attenuation Window (RAW) shows a 22% increase in the Average Satisfaction Ratio (ASR) for SUs. The simulations are conducted in the worst case scenario, where all SUs in the CRN are requesting to access the spectrum at the same time.

Introduction

Internet of Things (IoT) is a sophisticated paradigm that enables the communication among various physical objects – including sensors, actuators, household items, autonomous cars, mobile phones, and many others – through the Internet. Essentially, IoT managed to break the barrier between the physical and digital worlds bringing about tremendous improvements in several domains. From smart health care systems to smart cities where collected data from the sensors can be studied to manage sewage systems and power plants, the number of IoT services keeps rising daily [1]. In order to enable sustainable accessibility to remote services, IoT devices (IoDs) are interconnected through wireless communication protocols. However, the exploitation of wireless technologies – such as WiFi, Zigbee, and Bluetooth – to enable IoT communications is not a feasible solution, since most of these technologies cannot sustain coherent communications with the desired Quality of Service (QoS) [2], [3]. Furthermore, a prodigious rise in the number of IoDs to more than 30 billion is expected by the year 2025 [4]. Therefore, it is crucial to design wireless networks capable of handling IoDs without prompting spectrum scarcity.

Cognitive Radio Networks (CRNs) ensure efficient spectrum utilization by granting unlicensed Secondary Users (SUs) access to spectrum holes [5]. Therefore, it is desired for IoDs to be equipped with cognitive capabilities and act as SUs in CRNs [6], [7], [8], [9], [10], [11]. In [6], the authors discussed the importance of adding spectrum sensing to IoDs and presented challenges, such as security issues, that might face cognitive radio. The authors, in [7], discussed CRN-based IoT architectures and frameworks. Furthermore, they presented spectrum related functionalities in CRN-based IoT. In [4], the authors introduced frameworks for centralized and distributed AI-enabled IoT networks. Moreover, they analyzed multiple challenges for different network architectures. For instance, neural networks based approaches were employed to efficiently realize deep reinforcement learning strategies for spectrum access and spectrum sensing.

In CRNs, SUs sense the radio spectrum to utilize spectrum bands that are not occupied by Primary Users (PUs). Once a free spectrum band is detected, SUs can jointly access and use the vacant band. As the PU returns to the spectrum band, SUs must leave and utilize another idle spectrum band. From the preceding discussion, it is clear that spectrum sensing is the fundamental property of CRNs. Consequently, various spectrum sensing techniques have received a lot of attention from a large number of researchers [12]. Recently, the cooperative spectrum sensing approach was presented as an effective sensing method in CRNs [13], [14], [15], [16]. A model was proposed by [13] for cooperative spectrum sensing and access in CRNs using an overlapping coalitional game. In their work, SUs choose their cooperation strategy based on their corresponding traffic demand. Additionally, the SU utilizes an adaptive transmission power control scheme to further enhance its energy efficiency when a channel is assigned. In [15], the authors managed to model cooperative spectrum sensing in a CRN. In their study, SUs use an Inspection Game Based Learning (IGBL) algorithm to choose their cooperative strategy. Moreover, they developed a spectrum sharing scheme based on the Relative Utilitarian Bargaining Solution (RUBS). In [17], the authors studied a heterogeneous CRN, where K non-identical SUs and a fusion center collaboratively detect the PU’s signal using the N-out-of-K rule in the presence of erroneous control channels. Furthermore, they proposed a model to obtain the smallest number of SUs required in cooperative sensing while satisfying a target error bound at the fusion center.

For fair spectrum allocation, few interesting ideas were proposed [18], [19], [20], [21]. In [18], the authors developed a hybrid spectrum allocation approach whereby the spectrum is split between the macrocell and its nearby interfering femtocells based on their resource demands, while the far femtocells share the entire spectrum. They applied a lexicographic optimization procedure to decompose the formulated multi-objective problem into two subproblems and a low-complexity greedy algorithm was proposed to solve these subproblems sequentially. In [20], the authors developed a fair multi-channel assignment scheme for distributed cognitive radio networks. In addition, they integrated a new MAC framework for sensing and access contention resolution into their scheme. They devised the channel assignment problem according to Jain’s fairness criterion. Furthermore, they managed to achieve a good trade-off between throughput and fairness. In [21], the authors studied the fairness resource allocation problem in multi-user Rayleigh fading CRN with Co-Channel Interference (CCI) mitigation. They introduced the Correct Reception Probability (CRP) model as a metric for measuring the network utility. Their goal was to maximize the CRP of the worst performing SUs and control the CCI among SUs in each sub-channel. Their proposed algorithm for spectrum allocation showed geometrically fast convergence. Furthermore, the proposed power allocation algorithm guaranteed fairness for the throughput of different SUs in each sub-channel.

Despite of the extensive work done in cooperative spectrum sensing, the literature does not address the trade-off between intelligence and reliability of the systems. Owing to the development of AI-enabled intelligent SUs, there is no guarantee that SUs will follow the existing frameworks in case of abnormal events [22]. Therefore, a mechanism that enforces cooperation among Intelligent SUs must be established. Motivated by the preceding discussion, we propose a centralized CRN that primarily enforces cooperation and fairly assigns the spectrum among SUs. The main contributions of this paper are outlined as follows:

  • A Direct Strategy Selection (DSS) mechanism that enforces collaboration among SUs and satisfies the Aspiration Equilibrium (AE) criterion.

  • A novel spectrum allocation scheme based on the knapsack optimization problem, where the index of cooperation for each SU is taken into account for fair spectrum assignment.

  • Studying the effects of three windowing techniques on the index of cooperation and presenting the optimum window for the proposed scheme.

  • Exploring the trade-off between the utility function and cooperative behavior among SUs.

The rest of this paper is organized as follows. Section 2 introduces the centralized CRN architecture and formulates the mathematical model of the network. The proposed DSS mechanism and fair spectrum allocation scheme are presented in Section 3. Section 4 evaluates the performance of the proposed framework. Finally, Section 5 concludes the paper.

Section snippets

Network architecture

As shown in Fig. 1, the CRN consists of different clusters of intelligent IoDs (SUs). The Primary Base Station is defined as PBS, while the Cognitive Base Station is defined as CBS. The CBS governs the clusters. Each cluster monitors a PU, where SUs in each cluster sense the corresponding PU and jointly access its spectrum band in case of its absence. The spectrum access process starts with a channel estimation performed by each SU. Depending on the channel conditions, each SU determines its

Direct strategy selection

In this work, we adopt the notion of Aspiration Equilibrium (AE) for the agents to select their respective strategies. AE is an equilibrium state that could be found when an agent’s utility reaches or slightly exceeds its aspiration—hence the name. Based on this definition, we formulate Eq. (8) for Direct Strategy Selection (DSS). qa=max{Sa}w(Baαat),whereaAsa=min{Sa},qa<min{Sa}qa,min{Sa}qamax{Sa}max{Sa},qa>max{Sa}αat=αat1×δ(t),whereaAδ(t)=1tanh1(et×Tθ) where q is the calculated

Performance analysis

In this section, we thoroughly assess the proposed model which prioritizes cooperation and perform comparative analysis to a pre-existing model that prioritizes utility maximization. Furthermore, we test the models at the worst case scenario where all the SUs in a cluster request to access the spectrum at the same time. The parameters used for the simulations are summarized in Table 2 and the nodes are randomly distributes over an area of 1 km2 (1 km in X range and 1 km in Y range). Note that

Conclusion

In this paper, we presented a DSS mechanism that enforces cooperation among intelligent IoDs in CRNs. DSS satisfies the aspiration equilibrium, where each agent selects the strategy that will be most fit to satisfy its future aspiration. The algorithm countermeasures one of the main issues in cognitive systems that is the selfish behavior among intelligent agents. Moreover, we proposed a FSAS based on the knapsack optimization problem, where the amount of spectrum for each SU depends on its

CRediT authorship contribution statement

Ahmed H. Khalifa: Conceptualization, Methodology, Software, Writing - original draft, Writing - review & editing. Mohamed K. Shehata: Resources, Writing - review & editing. Safa M. Gasser: Resources, Writing - review & editing, Supervision. Mohamed S. El-Mahallawy: Writing - review & editing, Supervision.

Declaration of Competing Interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Ahmed Khalifa received the B.Sc. degree in Electronics and Communications Engineering from Arab Academy for Science, Technology and Maritime Transport, Egypt, in 2018. He is currently pursuing the M.Sc. degree in Computer Engineering. His research interests include computer vision and wireless communications.

References (26)

  • KhanA.A. et al.

    Cognitive radio-based internet of things: Applications, architectures, spectrum related functionalities, and future research directions

    IEEE Wirel. Commun.

    (2017)
  • ChiangM. et al.

    Fog and IoT: An overview of research opportunities

    IEEE Wirel. Commun.

    (2016)
  • ZhuR. et al.

    ERDT: Energy-efficient reliable decision transmission for intelligent cooperative spectrum sensing in industrial IoT

    IEEE Access

    (2015)
  • Cited by (0)

    Ahmed Khalifa received the B.Sc. degree in Electronics and Communications Engineering from Arab Academy for Science, Technology and Maritime Transport, Egypt, in 2018. He is currently pursuing the M.Sc. degree in Computer Engineering. His research interests include computer vision and wireless communications.

    Mohamed Shehata received the Ph.D. degree in information engineering Department of Electronics, Information and Bioengineering, Politecnico di Milano, Italy, in 2019. He is currently a lecturer at Arab Academy for Science and Technology and Maritime Transport—Faculty of Engineering, Cairo, Egypt. His research interests include 5G radio access networks, cloud networks, applications of machine learning to wireless networks, energy-efficiency and network resiliency.

    Safa Gasser received the Ph.D. degree in adaptive signal processing and control from the University of California at Santa Cruz, Santa Cruz, where she delivered the valedictory address at her graduation ceremony. She is currently an Assistant Professor of communications with the Department of Electronics and Communications, Arab Academy for Science, Technology and Maritime Transport (AASTMT), and a former Councilor of the IEEE AASTMT Chapter. Her research interests include adaptive signal processing, in addition to machine learning, remote sensing, and communications.

    Mohamed EL-Mahallawy (M’03) received the Ph.D. degree in image processing and pattern recognition from Cairo University, Egypt, in 2008, and the B.Sc. and M.Sc. degrees from the Electronics and Communications Department, Faculty of Engineering, Arab Academy for Science and Technology and Maritime Transport, Egypt, in 1998 and 2002, respectively. He was a Post-Doctoral Fellow with Universiti Technolgi Malaysia from 2011 to 2012. He is currently a Professor with the Electronics and Communications Department, Faculty of Engineering, Arab Academy for Science and Technology and Maritime Transport, Egypt.

    View full text