参与式感知系统中基于众包的动态群组构造方法研究

Participatory sensing technology and crowdsourcing applications have great potentials in areas of healthcare, enviroment protection, social life, etc.This project addresses crowdsourcing-based dynamic community creation approaches in participatory sensing systems.Three crowdsourcing models for participatory sensing were first proposed according to specific types of crowdsourcing applications. Community creation approaches will be developed based on each crowdsourcing model.The impacts of node mobiliy,social and opportunistic features on community creation are focused, and community creation schemes under opportunistic enviroment will be proposed as well. An efficient user reputation evaluation model is to be designed, which is further exploited in trustful community creation.User incentive polocies will be investigated based on data (or task) types, user reputation,and user profiles, and dynamic pricing gaming will be considered in community creation.Finally,we will build a participatory sensing experiment system with 100 smart phones and 100 sensor nodes for algorithm validation and performance evaluation.

参与式感知技术及其众包应用在公众健康、环境保护、社会生活等诸多领域都有着广阔的发展前景。本项目对参与式感知系统中基于众包的动态群组构造算法开展研究，首先根据参与式感知应用的特点提出三种众包模型，并且针对不同众包模型研究动态群组构造方法；研究参与式感知系统的移动性、社会特征和机会特征对群组构造的影响，以及在机会连接的背景下构造群组的方法；设计有效的用户声望评估模型，基于用户声望值进行可信群组构造；结合感知数据（或任务）类型、用户声望、用户档案上下文等要素设计用户激励策略，运用动态定价博弈方法进行群组构造。最后将构建一个由100个智能手机和100个传感器节点组成的参与式感知实验系统，并在实验系统上对设计的相关算法进行有效性验证和性能评价。

本项目对参与式感知系统中基于众包的动态群组构造算法开展研究。结合众包、群智感知、城市感知、人智感知等主题研究动态群组构造方法；研究参与式感知系统的移动性、社会特征和机会特征对群组构造的影响，研究了移动位置预测和用户移动轨迹分析，以及在机会连接的背景下构造群组的方法；研究群组构造中的激励与定价博弈，设计了基于节点依赖度、反向拍卖的节点激励方法，并以最大化系统收益为目标研究基于博弈论的群组构造和任务分配方法；结合移动通信网络的大数据，基于图论方法研究群组构造问题；研究了基于K-匿名、组合最优化等方法的用户位置隐私保护算法；构建了多个应用原型系统，包括空气质量监测感知系统、移动通信用户流失预测分析系统、基于移动大数据的商户选址推荐系统、空气质量感知的智跑APP、基于信号强度与邻区基站分析的用户定位APP、无线信号辐射值监测APP等。