群智感知中基于社会行为分析的机会式数据收集机制研究

With the development of society, the desire for thorough sensing is more and more strong. As a new sensing mode, crowd intelligence sensing has been a hot research area because of prominent advantages, such as wide coverage, low cost and strong scalability. As so far, the study of crowd intelligence sensing is still in the initial stage, and the theory and method of closely related research area are inefficient in such scenario. Focusing on three key problems in crowd intelligence sensing: social behavior abstract, participant recruitment and opportunistic data collection, this project research the abstract model and method for user behavior based on mobility profile, maps the actual behavior to calculable and verifiable virtual behavior space; We explore the mechanism of participant recruitment, design the recruitment framework based on social behavior analysis, and propose coverage and reputation based recruitment respectively; Based on the data collection mode which guided by the user behavior characteristics, low load distributed routing protocols and algorithms are proposed. Besides the theoretical proof and simulation, we also develop a social behavior analysis based haze monitoring crowd intelligence sensing system in order to test the efficiency of our study. The project concentrates on reducing data sensing cost, expanding the sensing coverage, increasing the user engagement and improving data quality and data collection performance in crowd intelligence sensing. The results of project will provide theoretical basis and technical support for development and popularization of crowd intelligence sensing applications.

社会发展到今天，对透彻感知的需求越来越强烈。群智感知作为一种覆盖范围广、成本低廉、扩展性强的新型感知模式，已经成为全球研究的热点。目前国内外对群智感知的研究尚处于起步阶段，而相关领域的理论和方法不适用于群智感知场景。本课题聚焦群智感知中社会行为抽象、参与者征募、机会式数据收集三大关键问题，重点研究基于移动性概要的用户行为抽象模型与方法，将用户实际行为映射到可计算、可验证的虚拟行为空间；探索参与者征募的机理，设计基于社会行为分析的参与者征募框架，提出基于覆盖和基于声誉的征募机制；建立以用户行为特征为指引的数据收集模式，提出低负载分布式路由协议及其相关算法；在理论证明和仿真测试的基础上，建立基于社会行为分析的雾霾监测群智感知系统。本项目致力于降低数据感知成本、增强感知时空覆盖度、提高用户参与度、提升数据质量和数据收集性能，为群智感知在更多新型应用领域的发展和普及提供理论基础和技术支撑。

本项目致力于降低数据感知成本、增强感知时空覆盖度、提高用户参与度、提升数据质量和数据收集性能，为群智感知在更多新型应用领域的发展和普及提供理论基础和技术支撑。本项目在以下四个方面取得了具有重要意义的研究结果：（1）提出了基于多行为属性的刻画方法和数据收集协议，在数据发送成功率和时延方面相对现有协议有改进。（2）提出了基于博弈理论的用户征募方法，可广泛应用于群智感知用户的征募中。（3）提出了一系列基于群智感知的数据机会式快速收集技术，促进数据的快速汇聚。（4）设计和开发了基于群智感知的应急搜救平台，推动应急搜救工作向数字化、智能化方向发展。