Crowdsensing and Resource Allocation in Shared Spectrum.
The exponential growth of mobile data services has translated into a proportionate surge in demand for greater wireless broadband capacity. Within today's xed spectrum allocation regime, exclusive spectrum access rights are granted to federal and commericial users, but a signicant portion of the licensed spectrum has been underutilized by primary users (PUs). To alleviate articial spectrum scarcity, spectrum sharing has been proposed to allow secondary users (SUs) to opportunistically access the locally unoccupied spectrum, called White Spaces (WS), so long as they do not cause harmful interference to PUs. To this end, the FCC is actively pursuing policy innovations to create shared spectrum, including WS in TV bands (TVWS) and the 3.5 GHz Citizens Broadcast Radio Service (CBRS) band, which often relies on a spectrum manager that manages the shared spectrum access, such as the database administrator (DBA) in TVWS and the Spectrum Access System (SAS) in CBRS. Our work begins by showing that the empirical DBA models for TV coverage estimation are locally inaccurate, since they do not explicitly account for local obstructions. Therefore, we propose augmenting the DBA approach with spatial-statistics-based radio mapping using Kriging and show that it achieves more accurate coverage boundary estimation, which leads to fewer missing WS opportunities (type-I errors) while keeping misclassications (type-II errors) under a certain limit. Scaling spatial-statistics-based radio mapping to larger areas inevitably meets cost limitations. An economically viable alternative is crowdsensing, that is, outsourcing sensing tasks to spatially distributed users with mobile devices that are outtted with spectrum sensors. In order to attract user participation for crowdsensing, we propose an auction-based incentive mechanism, in which each user submits a bid (the minimum acceptable payment) for providing spectrum data and receives a payment when selected. We show that the proposed scheme is truthful, computationally ecient, individually rational, and budget feasible. We also consider the design of a pricing-based incentive mechanism, in which the platform who constructs radio maps makes one-time oers (the incentive for participation) to selected users (either sequentially or in batches) and collects data from those who accept the offers. We formulate pricing mechanism design as expected utility maximization, where the expected utility captures the tradeo between radio mapping performance (location and data quality), crowdsensing cost, and uncertainty in oer outcomes (possible expiration and rejection). We show that the proposed user selection algorithm provides a provable performance guarantee and the proposed mechanism outperforms the baseline mechanisms. After WS opportunities are identied, it is crucial to eciently allocate resources (e.g., available channels) to SUs. To this end, we study SAS-assisted dynamic channel assignment in the CBRS. We propose a novel graph representation to capture spatially varying channel availability,…
Advisors/Committee Members: Roy, Sumit (advisor), Poovendran, Radha (advisor).
to Zotero / EndNote / Reference
APA (6th Edition):
Ying, X. (2018). Crowdsensing and Resource Allocation in Shared Spectrum. (Doctoral Dissertation). University of Washington. Retrieved from http://hdl.handle.net/1773/42294
Chicago Manual of Style (16th Edition):
Ying, Xuhang. “Crowdsensing and Resource Allocation in Shared Spectrum.” 2018. Doctoral Dissertation, University of Washington. Accessed October 19, 2018.
MLA Handbook (7th Edition):
Ying, Xuhang. “Crowdsensing and Resource Allocation in Shared Spectrum.” 2018. Web. 19 Oct 2018.
Ying X. Crowdsensing and Resource Allocation in Shared Spectrum. [Internet] [Doctoral dissertation]. University of Washington; 2018. [cited 2018 Oct 19].
Available from: http://hdl.handle.net/1773/42294.
Council of Science Editors:
Ying X. Crowdsensing and Resource Allocation in Shared Spectrum. [Doctoral Dissertation]. University of Washington; 2018. Available from: http://hdl.handle.net/1773/42294