Multiagent Reinforcement Learning Based Spectrum Sensing Policies for Cognitive Radio Networks

Lundén, Jarmo; Kulkarni, Sanjeev R.; Koivunen, Visa; Poor, H. Vincent

doi:10.1109/jstsp.2013.2259797

Cited by 59 publications

(48 citation statements)

References 24 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Ref. [38] aims to integrate HetNets by reinforcement learning with limited feedback and noisy information. Game theory based learning and sequential decision making are also investigated in Refs.…”

Section: Incompleteness and Inconsistencymentioning

confidence: 99%

Wireless big data: transforming heterogeneous networks to smart networks

Huang

Tan

Liang

2017

J. Commun. Inf. Netw.

View full text Add to dashboard Cite

In HetNets (Heterogeneous Networks), each network is allocated with fixed spectrum resource and provides service to its assigned users using specific RAT (Radio Access Technology). Due to the high dynamics of load distribution among different networks, simply optimizing the performance of individual network can hardly meet the demands from the dramatically increasing access devices, the consequent upsurge of data traffic, and dynamic user QoE (Quality-of-Experience). The deployment of smart networks, which are supported by SRA (Smart Resource Allocation) among different networks and CUA (Cognitive User Access) among different users, is deemed a promising solution to these challenges. In this paper, we propose a framework to transform HetNets to smart networks by leveraging WBD (Wireless Big Data), CR (Cognitive Radio) and NFV (Network Function Virtualization) techniques. CR and NFV support resource slicing in spectrum, physical layers, and network layers, while WBD is used to design intelligent mechanisms for resource mapping and traffic prediction through powerful AI (Artificial Intelligence) methods. We analyze the characteristics of WBD and review possible AI methods to be utilized in smart networks. In particular, the potential of WBD is revealed through high level view on SRA, which intelligently maps radio and network resources to each network for meeting the dynamic traffic demand, as well as CUA, which allows mobile users to access the best available network with manageable cost, yet achieving target QoS (Quality-of-Service) or QoE.

show abstract

Section: Incompleteness and Inconsistencymentioning

confidence: 99%

Wireless big data: transforming heterogeneous networks to smart networks

Huang

Tan

Liang

2017

J. Commun. Inf. Netw.

View full text Add to dashboard Cite

show abstract

“…The capture threshold and the collision detection threshold were both set to 7 dB. The SUs employ collaborative energy detection and the multiagent reinforcement learning based sensing policy proposed in [5]. In this sensing policy the SUs employ collaboration to maximize the amount of vacant spectrum found.…”

Section: E Example: Reinforcement Learning Based Sensing Policy Withmentioning

confidence: 99%

“…In this sensing policy the SUs employ collaboration to maximize the amount of vacant spectrum found. The idea in [5] is to obtain a balance between sensing different PU channels and thus sensing more spectrum and increasing the sensing reliability by sensing the same PU channels. Sensing time and probability of detection are kept fixed.…”

Section: E Example: Reinforcement Learning Based Sensing Policy Withmentioning

confidence: 99%

“…The higher the number of SUs sensing simultaneously the same PU channel, the smaller the probability of false alarm and thus the smaller the number of overlooked spectral opportunities as well. The employed reinforcement learning algorithm and energy detection parameters are the same as in [5]. Note that in this spectrum sensing scheme the SUs make distributed decisions which frequency bands to sense.…”

Section: E Example: Reinforcement Learning Based Sensing Policy Withmentioning

confidence: 99%

“…Such algorithms and/or protocols have utility, in particular, in cognitive radio ad hoc networks [4] in which there is no centralized controller or base station to control the sensing operations. An example of such a cooperative sensing system has been proposed in [5].…”

Section: Introductionmentioning

confidence: 99%

See 2 more Smart Citations

Distributed Algorithms for Sharing Spectrum Sensing Information in Cognitive Radio Networks

Lundén

Motani

Poor

2015

IEEE Trans. Wireless Commun.

Self Cite

View full text Add to dashboard Cite

Collaborative spectrum sensing in cognitive radio networks mitigates the negative propagation effects of the wireless channel and increases the sensing reliability. Collaborative sensing requires sensors to share their local spectrum sensing information (SSI) with other users. This paper proposes distributed iterative time slot allocation algorithms for SSI sharing on a dedicated common control channel in a cognitive radio ad hoc network scenario. The proposed algorithms are based on a collision detection and acknowledgment scheme. This scheme allows the network nodes to receive knowledge about collisions regarding their transmitted SSI packets. The nodes use this information to update their operating time slots using a probabilistic approach; each node maintains and updates a parameter representing the probability of switching the time slot in case of a collision. Both fixed and adaptive probability based schemes are proposed. The proposed algorithms are proven to converge to a collision-free allocation with probability one if such an allocation exists. Moreover, an analytical expression for the expected convergence time is established. Extensive simulation results illustrating the rapid convergence, excellent performance, and small reporting overhead of the proposed time slot allocation algorithms are provided.Index Terms-Cognitive radio ad hoc networks, common control channel, distributed time slot allocation, spectrum sensing information sharing.

show abstract

Machine learning for cooperative spectrum sensing and sharing: A survey

Janu

Singh

Kumar

2021

Trans Emerging Tel Tech

View full text Add to dashboard Cite

With the rapid development of next-generation wireless communication technologies and the increasing demand of spectrum resources, it becomes necessary to introduce learning and reasoning capabilities in cognitive radio networks (CRN). In particular, our focus is on two fundamental applications in CRNs, namely spectrum sensing (SS) and spectrum sharing. The application of machine learning (ML) techniques has added new aspects to SS and spectrum sharing. This paper offers a survey on various ML-based algorithms in the cooperative spectrum sensing (CSS) and dynamic spectrum sharing (DSS) domain, with its emphasis on types of features extracted from primary user signal, types of ML algorithm, and performance metrics utilized for evaluation of ML algorithms. Starting with the basic principles and challenges of SS, this paper also justifies the applicability of supervised, unsupervised, and reinforcement ML algorithms in the CSS domain. The application of ML algorithms, to solve the DSS problem has also been reviewed. Finally, the survey paper is concluded with some suggested open research challenges and future directions for ML application in next-generation communication technologies.

show abstract

Multiagent Reinforcement Learning Based Spectrum Sensing Policies for Cognitive Radio Networks

Cited by 59 publications

References 24 publications

Wireless big data: transforming heterogeneous networks to smart networks

Wireless big data: transforming heterogeneous networks to smart networks

Distributed Algorithms for Sharing Spectrum Sensing Information in Cognitive Radio Networks

Machine learning for cooperative spectrum sensing and sharing: A survey

Contact Info

Product

Resources

About