NOn-parametric Bayesian channEls cLustering (NOBEL) Scheme for Wireless Multimedia Cognitive Radio Networks
- Authors
- Ali, Amjad; Ahmed, Muhammad Ejaz; Ali, Farman; Tran, Nguyen H.; Niyato, Dusit; Pack, Sangheon
- Issue Date
- 10월-2019
- Publisher
- IEEE-INST ELECTRICAL ELECTRONICS ENGINEERS INC
- Keywords
- Wireless multimedia applications; multimedia CRNs; multi-channel; channel clustering; QoS-level quantification
- Citation
- IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS, v.37, no.10, pp.2293 - 2305
- Indexed
- SCIE
SCOPUS
- Journal Title
- IEEE JOURNAL ON SELECTED AREAS IN COMMUNICATIONS
- Volume
- 37
- Number
- 10
- Start Page
- 2293
- End Page
- 2305
- URI
- https://scholar.korea.ac.kr/handle/2021.sw.korea/62711
- DOI
- 10.1109/JSAC.2019.2933943
- ISSN
- 0733-8716
- Abstract
- In wireless multimedia cognitive radio networks (WMCRNs), to optimize multimedia transmissions and scarce wireless spectrum utilization, a multimedia secondary user (MSU) needs to estimate and/or identify the achievable quality of service (QoS)-levels over the available licensed channels. However, due to the lack of signaling information among MSUs and the primary users (PUs) in uncoordinated environments, identification of the achievable QoS-levels on the available licensed channels is a challenging problem and has not yet been fully explored. To address this challenge, we propose a novel NOn-parametric Bayesian channEls cLustering (NOBEL) scheme. In NOBEL, an infinite Gaussian mixture model-based collapsed Gibbs sampler is adopted to identify the achievable QoS-levels over the feature space, i.e., bitrate, packet delay variation, and packet delivery ratio on the PUs' licensed channels. Real trace-driven evaluation results demonstrate that NOBEL outperforms other baseline clustering techniques and guarantee high accuracy from 98% to 99.5%.
- Files in This Item
- There are no files associated with this item.
- Appears in
Collections - College of Engineering > School of Electrical Engineering > 1. Journal Articles
Items in ScholarWorks are protected by copyright, with all rights reserved, unless otherwise indicated.