Mobile social network analytics and mobile edge solutions for trustworthy and reliable urban sensing

越来越需要智能方法，以便能够快速可靠地感知、分析和呈现有关应急准备、公共安全、公共卫生、环境和城市环境中的生活质量的信息。需要这些方法来改善城市感知并确保智能人群感知基础设施。示例包括智能手机用户的隐性招募，以感知交通状况、环境污染、噪声水平或温度，在某些位置捕获图像，并快速分析收集的数据，以便为公民和/或当局准备即时报告。然而，在人群感知中，数据源是嘈杂的，不可靠的，错误的，并且在很大程度上是未知的。在我们以前的工作中，我们通过拍卖理论和博弈论的方法解决了用户招募问题，其目标是最大限度地提高被招募用户集在众测应用中的可信度。我们已经证明，使用统计和协作声誉系统可以显着提高所获得的数据的有用性。尽管人群感知研究取得了进展，但仍然缺乏一个可靠，安全，可扩展和强大的框架，可用于有效地收集，量化和分析数据。由于任何人都可以在下载移动的应用程序后参与城市感测，因此没有彻底检查源池，并且数据收集器通常不知道单个参与者的可靠性。因此，特别是在公共安全或应急准备等使命关键应用中，提炼可靠信息的问题变得具有挑战性。该研究计划的目标是为城市环境中基于移动的社交网络的数据采集奠定基础，并解决其可靠性，可扩展性和隐私方面的挑战。事实上，移动的社交网络表示用户之间的交互链接，其可以是各种形式，诸如数据通信、共址和移动性。该研究将利用网络科学，数据分析，建模和优化以及多学科分析和设计领域的方法和结果。研究结果将对社交网络辅助的城市传感系统产生变革性影响，特别是在可持续性和公共安全领域，因为它具有以下新的组成部分：1）用于有效用户招募的连续社会识别和认证算法，2）用于提高可靠性的新颖数据融合算法，3）雾计算架构的可扩展实现，以处理意外增加的人群感测数据的量和速度，4）隐藏上下文模式以提高参与者的隐私。研究结果将导致一个新的框架，包括社会网络驱动的城市传感系统提供可扩展性，可信性和隐私。

There is a growing need for smart methodologies to enable rapid and reliable sensing, analysis and presentation of information regarding emergency preparedness, public safety, public health, environment and quality of life in urban environments. These methodologies are needed to improve urban sensing and ensure a smart crowdsensing infrastructure. Examples include implicit recruitment of smartphone users to sense traffic conditions, environmental pollution, noise levels or temperature, capture images at certain locations, and analyze collected data rapidly so that immediate reports can be prepared for citizens and/or authorities. However, in crowdsensing, data sources are noisy, unreliable, erroneous, and largely unknown. In our previous work, we addressed the user recruitment problem via auction-theoretic and game-theoretic approaches with the objective of maximizing trustworthiness of the recruited user set in crowdsensing applications. We have shown that the use of statistical and collaborative reputation systems can significantly improve the usefulness of acquired data. Despite the progress in crowdsensing research, a reliable, secure, scalable and robust framework that can be used to effectively collect, quantify and analyze data is lacking. As anyone is allowed to participate in urban sensing upon downloading a mobile app, the pool of sources is not examined thoroughly, and the reliability of individual participants is generally unknown to the data collector. Hence, particularly in mission critical applications such as public safety or emergency preparedness, the problem of refining reliable information becomes challenging. The objective of this research program is to lay the foundations for mobile social network-based data acquisition in urban settings and address its reliability, scalability, and privacy challenges. Indeed, mobile social network denotes interaction links between users, which can be in various forms such as data communications, co-location and mobility. The research will leverage methods and results from the fields of network science, data analytics, modelling and optimization, and multidisciplinary analysis and design. The research findings will be transformative for social network-assisted urban sensing systems, especially in sustainability and public safety areas owing to the following novel components: 1) continuous social identification and authentication algorithms for effective user recruitment, 2) novel data fusion algorithms to improve reliability, 3) scalable implementation of fog computing architecture to handle unexpectedly increasing volume and velocity of crowd-sensed data, 4) cloaked contextual patterns to improve privacy of the participants. The research findings will lead to a novel framework consisting of social network-driven urban sensing system offering scalability, trustworthiness and privacy.