基于混合群智感知的位置服务自适应提供机理研究

It is the core of intelligent services to improve the comprehensive adaptability of the provisioning mechanism for LBS (Location-based services). This project will deeply analyse the mobile and social characteristics of LBS, and explore the organization of the semantic location, the QoE evaluation and the description model which can support the expression of the context, to improve the sharing adaptability of the service. The abstract method of the service and the mechanism of the fault tolerance via Self-evolvement and the sleeping via Life Game for the internal nodes are researched, to enhance the robustness of the service. Then the data transmission mechanism based on the situation awareness in the Hybrid Crowdsensing, and the fusion method of the contextual data based on distributed rules and the community detection respectively are explored, also the updating strategy of the context is introduced, to obtain the individual and the group context characteristics of the user, to improve the context adaptability of the service. An open service discovery framework was constructed according to the capability-driven method, by introducing the virtual entity and the adaptive localization model. And the service selection method via the local approximate filtering based on QoE is adopted in the framework, to improve the requirement adaptability of the service discovery. In the implementation of the service, the checkpoint Agent and the environment-driven model are introduced to achieve the monitoring and handling with the exception caused by both the user's needs and the service resources. For the invalid service, the service replacement mechanism with compensation-support based on transaction by considering the location update was presented, to enhance the exception adaptability of the service implementation. The outcomes delivered by this project will provide theoretical guidance and practical methods for the construction of the service system of LBS, and lay the foundation for pushing location-aware computing of LBS.

增强位置服务提供的整体适应性，是智能化服务的核心。本项目通过深入分析位置服务的移动和社会化特征，探索其语义位置组织、QoE评价、以及情境可表达的描述模型，提高服务的共享适应性；研究服务的抽象及其内部自主演化的容错和仿“生命游戏”的休眠机制，增强服务的鲁棒性；探索混合群智感知中态势感知的数据传输、和基于分布式规约和社区发现的情境数据融合方法，并引入情境更新策略发掘用户的个/群体综合情境特征，增强服务的情境适应性；采用能力驱动方法，引入虚拟实体和自适应定位模型，构建开放的服务发现框架，并通过QoE局部近似过滤进行服务选择，提高服务发现的需求适应性；在服务执行过程中，引入检查点Agent和环境驱动模型实现用户需求和服务资源双异常的监控处理，针对失效服务，探索考虑重定位的事务级可补偿服务替换方法，增强服务执行的异常适应性。研究成果将为位置服务体系构建提供理论和实验依据，为推动位置服务计算奠定基础。

如何将用户需求与位置服务更好地适配，是智能化服务的核心任务，因此本项目探索了增强位置服务共享适应性、鲁棒性、情境适应性、需求适应性以及异常适应性的一系列方法。首先围绕位置服务的语义共享机制开展研究，深入分析了位置的社会性信息，提出了基于本体的语义位置信息模型、可扩展的多层次QoE量化模型、基于QoE的位置服务语义描述模型；其次探索了混合群智感知环境下感知节点的数据采集方法，提出了考虑数据质量和覆盖度的混合群智感知信息融合方法，以及面向群智感知的高效任务分配、调度和定价方法；再次，项目组围绕位置服务的自适应发现开展研究，设计了开放的服务发现框架，研究了高精度、高效的室内定位方法，构建了情境感知的自适应定位模型，以及提出了面向QoE约束的基于人工鱼群和蚁群算法的服务选择算法；最后，项目组围绕位置服务的异常处理开展研究，提出了基于检查点Agent的用户需求异常处理模型，基于环境感知的服务资源异常处理模型，以及考虑用户兴趣的服务替换方法。本项目已发表学术论文22篇，包括期刊论文15篇（SCI收录）、会议论文7篇（EI收录），获批发明专利2项，参加学术交流活动19人次，组织国际学术会议2次，共培养14名博/硕士研究生。该项目的研究成果不仅可以为位置服务体系构建提供理论和实验依据，还可以为推动位置服务计算奠定基础。