移动群智感知中的任务优化分配模型与方法研究

Mobile crowdsensing fully utilizes the wide mobility of smart devices, typically equipped with multiple types of sensors, to realize wide range and long-term sensing coverage of physical environments with multi-dimensional sensing information collection. It has the advantages of low cost, high dynamics, and good scalability. In this project, we start with the high diversity of user and task distribution, consider limitation of user resources and budgets, task quality characteristics, and user participation modes, conduct in-depth research on modeling and method design for task assignments in mobile crowdsensing systems. We focus on three key research problems: how to achieve optimized task assignments subject to personalized resource constraints, how to achieve optimized task assignments by jointly utilizing opportunistic sensing and participatory sensing, and how to optimize participating groups. The major research contents in this project are as follows: research on task assignment methods using opportunistic sensing subject to constraints of communication opportunities, residual energy; research on task assignment methods using participatory sensing subject to traveling distance constraint; and research on task-, data-, or user-centric joint design of task assignment methods by considering both opportunistic sensing users and participatory sensing users, incentive mechanisms for selecting participating user candidates based on task properties and distributions. In this research, we shall design a series of novel and effective task assignment methods by considering different resource limitations, application scenarios, and user participating modes, which are expected to greatly improve task quality and reduce system costs while enriching the theory and key techniques of mobile crowdsensing field.

移动群智感知充分利用大量配备多种传感器的智能便携设备的时空移动性，实现对物理世界的大范围、长时间的深度覆盖和广泛的多维信息采集，具有成本低、动态性强和扩展性好的优势。本项目拟针对移动群智感知的多样性应用场景，从用户和任务的时空分布特性出发，结合用户资源、预算约束、任务质量、参与模式等因素，围绕用户个性化资源约束下的任务优化分配、主动和机会感知联合任务优化分配和参与用户群体组成优化三个关键问题，构建针对不同目标的优化模型，设计相应的任务优化分配方法和激励机制，主要包括：研究通信费用、电量水平等约束下的机会感知任务分配方法；移动约束下的主动感知任务分配方法；以任务、数据或用户为中心的主动与机会感知相协同的联合任务分配方法；结合任务特性及分布的用户群体优化的激励机制。研究过程中将设计一系列针对不同资源条件和设计目标的优化模型与方法，显著提升任务质量、降低系统成本，丰富移动群智感知的理论和方法。

移动群智感知是物联网研究与应用的热点问题，在室内定位、污染检测、交通信息采集和公共安全等领域具有广泛的应用前景。移动群智感知可以充分利用配备多种传感器的智能便携设备的移动性，实现对物理世界的大范围、长时间的深度覆盖和广泛的多维信息采集，具有成本低、动态性强和扩展性好的优势。本项目针对移动群智感知的多样性应用场景，从用户和任务的时空分布特性出发，结合用户资源、任务质量、参与模式等因素，构建针对不同目标的优化模型，展开了以下研究：群智感知任务分配与激励机制、移动边缘计算的任务调度和资源分配等。课题研究过程中，设计了一系列针对不同应用场景要求的新型任务分配算法和激励机制，显著提高任务完成效率、降低系统花费和增加社会效益。课题研究过程中，取得了一些列研究成果，共发表学术论文44篇，其中国际学术刊物29篇（其中IEEE期刊19篇，包括IEEE Transactions on Wireless Communications、IEEE/ACM Transactions on Networking、 IEEE Transactions on Vehicular Technology、IEEE Network、IEEE Wireless Communications、IEEE Internet of Things Journal等）、国际学术会议15篇（其中IEEE通信学会旗舰会议11篇），申请国家发明专利4项（授权1项），毕业研究生8名（博士生6名、硕士生2名），其中北京市优秀毕业生2名，为群智感知的发展提供了技术支持、培养了人才。