西安电子科技大学学报 ›› 2022, Vol. 49 ›› Issue (1): 142-151.doi: 10.19665/j.issn1001-2400.2022.01.014

• 信息与通信工程 • 上一篇    下一篇

偏好感知的边云协同群智感知参与者选择策略

王汝言1,2,3(),刘佳1,2,3(),何鹏1,2,3(),崔亚平1,2,3()   

  1. 1.重庆邮电大学 通信与信息工程学院,重庆 400065
    2.重庆高校市级光通信与网络重点实验室,重庆 400065
    3.泛在感知与互联重庆市重点实验室,重庆 400065
  • 收稿日期:2020-11-28 出版日期:2022-02-20 发布日期:2022-04-27
  • 通讯作者: 何鹏
  • 作者简介:王汝言(1969—),男,教授,博士,E-mail: wangry@cqupt.edu.cn;|刘 佳(1996—),男,重庆邮电大学硕士研究生,E-mail: 1747543987@qq.com;|崔亚平(1986—),男,讲师,博士,E-mail: cuiyp@cqupt.edu.cn
  • 基金资助:
    国家自然科学基金(61901070);国家自然科学基金(61871062);国家自然科学基金(61771082);国家自然科学基金(61801065);重庆市教委科学技术研究项目(KJQN201900611);重庆市教委科学技术研究项目(KJQN202000603);重庆市高校创新研究群体(CXQT20017);重庆市自然科学基金重点项目(cstc2020jcyj-zdxmX0024)

Preference aware participant selection strategy for edge-cloud collaborative crowdsensing

WANG Ruyan1,2,3(),LIU Jia1,2,3(),HE Peng1,2,3(),CUI Yaping1,2,3()   

  1. 1. School of Communication and Information Engineering,Chongqing University of Posts and Telecommunications,Chongqing 400065,China
    2. Chongqing Key Laboratory of Optical Communication and Networks,Chongqing 400065,China
    3. Chongqing Key Laboratory of Ubiquitous Sensing and Networking,Chongqing 400065,China
  • Received:2020-11-28 Online:2022-02-20 Published:2022-04-27
  • Contact: Peng HE

摘要:

群智感知依靠大量用户的移动性和智能设备的传感能力完成数据的收集,已经成为一种有效的感知数据收集方式。现有集中式群智感知网络中,云平台负责任务分发和数据收集等全部过程,难以有效处理大量实时数据,感知成本高;不同参与者对待任务的兴趣不同,忽略偏好因素会导致所选参与者完成任务的效率较低,参与者满意度较差。针对上述群智感知网络中存在的问题,提出一种边云协同架构下偏好感知的参与者选择策略。参与者选择过程由云平台和边缘节点协作执行。云平台基于任务的不同位置向边缘节点分发任务,并且从边缘节点收集数据。边缘节点负责参与者选择过程,通过评估时间匹配度、距离匹配度、任务类型和奖励,量化用户对任务的偏好;通过评估用户的声誉和感知成本,量化任务对用户的偏好。基于双方偏好和稳定匹配理论,将参与者选择问题建模为用户与任务间的多对一稳定匹配问题,并且求解稳定匹配使参与者偏好最大化。结果表明,该策略所选参与者的满意度较高,平台收集的数据质量较好。

关键词: 群智感知, 边云协同, 参与者选择, 数据质量, 用户偏好

Abstract:

Crowdsensing relies on the mobility of a large number of users and the sensing ability of intelligent devices to complete data collection,which has become an effective way of sensing data collection.In the existing crowdsensing network,the cloud platform is responsible for the whole process of task distribution and data collection,so it is difficult to effectively process a large amount of real-time data and the sensing cost is high.Different participants have different interests in tasks.Ignoring preference factors will lead to a low efficiency and poor satisfaction of selected participants.Aiming at the problems existing in the above crowdsensing network,a preference aware participant selection strategy under the edge-cloud collaborative architecture is proposed.The participant selection process is performed by the cloud platform and edge nodes in collaboration.The cloud platform distributes tasks to edge nodes based on different locations of tasks,and collects data from edge nodes.The edge node is responsible for the participant selection process,quantifying the user's preference for the task by evaluating the time matching degree,distance matching degree,task type and reward,and quantifying the user's preference for the task by evaluating the user's reputation and sensing cost.Based on the bilateral preference and stable matching theory,the participant selection problem is modeled as a many-to-one stable matching problem between users and tasks,with the stable matching solved to maximize the participant preference.The results show that the participants selected by this strategy have high satisfaction and that the data quality collected by the platform is good.

Key words: crowdsensing, edge cloud collaboration, participant selection, data quality, user preference

中图分类号: 

  • TN929.5