Crowd-Empowered Privacy-Preserving Data Aggregation for Mobile Crowdsensing
Crowd-Empowered Privacy-Preserving Data Aggregation for Mobile Crowdsensing
复制标题
DOI：
10.1145/3209582.3209598
复制
发表时间：
2018-06
期刊：
影响因子：
--
通讯作者：
Lei Yang;Mengyuan Zhang;Shibo He;Ming Li;Junshan Zhang
中科院分区：
文献类型：
--
作者：
Lei Yang;Mengyuan Zhang;Shibo He;Ming Li;Junshan Zhang
我们开发了一个拍卖框架，隐私保护数据聚合在移动的crowdsensing，该平台扮演的角色，作为一个拍卖师招募工人的传感任务。在这个框架中，工作人员被允许报告其数据的隐私保护版本，以保护他们的数据隐私;平台根据他们的感知能力选择工作人员，旨在解决由于存在多个纳什均衡而无法确保聚合结果准确性的博弈论模型的缺点。观察到，在这种基于拍卖的框架中，在工作者的数据隐私之间存在外部性，因为每个工作者的数据隐私取决于她注入的噪声和聚合结果中的总噪声，该总噪声与选择哪些工作者来完成任务密切相关。为了以具有成本效益的方式实现数据聚合的理想准确性水平，我们明确描述了外部性，即，每个工作人员添加的噪声对数据隐私和聚合结果的准确性的影响。进一步，我们探索了问题的结构，刻画了问题的隐单调性，并确定了工人的临界出价，这使得设计一个真实的，个人理性的和计算效率高的激励机制成为可能。所提出的激励机制可以招募一组工人，以近似最小化从工人那里购买私人传感数据的成本，但要满足聚合结果的准确性要求。我们通过理论分析和大量的仿真验证了所提出的方案。
We develop an auction framework for privacy-preserving data aggregation in mobile crowdsensing, where the platform plays the role as an auctioneer to recruit workers for a sensing task. In this framework, the workers are allowed to report privacy-preserving versions of their data to protect their data privacy; and the platform selects workers based on their sensing capabilities, which aims to address the drawbacks of game-theoretic models that cannot ensure the accuracy level of the aggregated result, due to the existence of multiple Nash Equilibria. Observe that in this auction based framework, there exists externalities among workers' data privacy, because the data privacy of each worker depends on both her injected noise and the total noise in the aggregated result that is intimately related to which workers are selected to fulfill the task. To achieve a desirable accuracy level of the data aggregation in a cost-effective manner, we explicitly characterize the externalities, i.e., the impact of the noise added by each worker on both the data privacy and the accuracy of the aggregated result. Further, we explore the problem structure, characterize the hidden monotonicity property of the problem, and determine the critical bid of workers, which makes it possible to design a truthful, individually rational and computationally efficient incentive mechanism. The proposed incentive mechanism can recruit a set of workers to approximately minimize the cost of purchasing private sensing data from workers subject to the accuracy requirement of the aggregated result. We validate the proposed scheme through theoretical analysis as well as extensive simulations.