Dissertation/Thesis Abstract

Ultra-Large-Scale Crowdsensing in Device-to-Device Networks
by Han, Yanyan, Ph.D., University of Louisiana at Lafayette, 2017, 120; 10266292
Abstract (Summary)

Crowdsourcing is emerging as a new data-collection, solution-finding, and opinion-seeking model that obtains needed services, ideas, or content by soliciting contributions from a large crowd of public participants. D2D based crowdsensing is particularly desired when the initiator cannot directly reach out to the participants or the conventional approaches for data transportation are costly. This dissertation studies the ultra-large scale crowdsensing applications in such mobile D2D networks. First, I proposed and addressed MCC(Minimum-Cost Crowdsourcing) problem by exploring a multi-dimensional design space to seek an optimal solution that minimizes the total crowdsensing cost while satisfying the coverage probability over the FoI. In particular, three strategies (or options) are in consideration: task allocation strategy, data processing strategy and computation offloading strategy. The difficulty is to determine the three options for each node in order to minimize the overall system cost. Second, there are a class of applications, where the originator is only allowed to recruit a given number of participants. Therefore, from the perspective of limited participants, we proposed a competition based participant recruitment mechanism to wisely choose the set of nodes while achieving the best benefit. I have proposed a dynamic programming algorithm as a first attack to this problem, followed by two distributed alternatives, which prove to be more practical and adaptive. During the above two topics, we find the existing routing protocols cannot efficiently support the ultra-large scale crowdsensing, thus we built a resource constrained routing protocol in D2D, aiming to approach the large-scale, bandwidth-hungry crowdsensing task in a more efficient way. With the requirement of restricted node storage and link bandwidth as well as end-to-end delay, I formulated a non-linear traffic allocation optimization problem with an approximation algorithm and distributed heuristic solution. Finally, I have carried out extensive complexity analysis, simulation, prototyping and implementation, experimentation and performance evaluation. Through the step-by-step exploration and verification, I have demonstrated the efficiency of the proposed heuristics and revealed empirical insights into the design tradeoffs and practical considerations in D2D-based crowdsourcing.

Indexing (document details)
Advisor: Wu, Hongyi
Commitee: Bayoumi, Magdy A., Perkins, Dimitri, Tzeng, Nian-Feng, Wu, Hongyi
School: University of Louisiana at Lafayette
Department: Computer Science
School Location: United States -- Louisiana
Source: DAI-B 79/01(E), Dissertation Abstracts International
Source Type: DISSERTATION
Subjects: Computer science
Keywords: Crowdsensing, Device-to-device network, Opportunistic, Optimization, Prototype
Publication Number: 10266292
ISBN: 978-0-355-11416-4
Copyright © 2019 ProQuest LLC. All rights reserved. Terms and Conditions Privacy Policy Cookie Policy
ProQuest