面向复杂关联场景下状态时变实体的高效搜索技术研究

The Internet of Things (IoT) search technology plays a vital role in promoting the sharing and applications of sensing resource services and data, and advancing the deep integration of physical world and information space. In view of the complex association relationships between entities and sensing devices and the spatiotemporal dynamics of entity states, aimming at improving the search efficiency towards IoT entities in the complex association scenarios, according to the research route of search cooperating among sensing devices, differentially caching for entity contents, and search matching towards physical entities, this project carries out research including the search collaborating technology among sensing devices with complex associations, differentiated caching technology for contents on entities with time-varying states, and entity matching technology with observation content prediction. The specific research objectives are as follows. We propose to break through the sensing device search collaboration technology and solve the problem of ordered coordination among multi-associative sensing devices in the complex one-to-many association scenarios. Besides, the observation content caching mode for IoT entities with time-varying states is planed to be innovated to alleviate the efficiency problem caused by traditional undifferentiated caching pattern. This project also aims to improve the entity search matching mechanism with observation content prediction by constructing high-precision and more universal observation content prediction model in order to enhance the accuracy of entity matching mechanism based on observation contents. The research results of this project will promote the fundamental theoretical innovations and large-scale applications in the field of IoT search.

物联网搜索技术对推动感知资源服务与数据的共享及应用，促进物理空间与信息空间的深度融合具有至关重要的作用。针对实体与传感设备关联关系的复杂性及实体状态的时空动态性特征，本项目以提高面向复杂关联场景下物联网实体的搜索效率为目标，按照“传感设备搜索协同→实体内容分类缓存→物理实体搜索匹配”的研究路线，拟开展复杂关联场景下传感设备搜索协同、面向状态时变实体内容的分类缓存、带有观测内容预测的实体搜索匹配三个方面内容的研究。具体目标如下：拟突破复杂关联场景下传感设备搜索协同技术，解决“一对多”复杂关联场景下传感设备间的有序协同问题；拟创新差异化状态时变实体的观测内容缓存模式，解决传统无差别缓存模式效率较低的问题；拟改进观测内容预测的实体搜索匹配机制，构建高精度、高普适性的观测内容预测模型，以提高基于观测内容的实体匹配机制的精度。本项目的研究对推动物联网搜索领域的基础理论创新与大规模应用具有促进意义。

物联网搜索技术对推动感知资源服务与数据的共享及应用，促进物理空间与信息空间的深度融合具有至关重要的作用。针对实体与传感设备关联关系的复杂性及实体状态的时空动态性特征，本项目以提高面向复杂关联场景下物联网实体的搜索效率为目标，按照“传感设备搜索协同→实体内容分类缓存→物理实体搜索匹配”的研究路线，开展了相关研究工作。并在此基础上，进一步扩展了隐私实体数据搜寻、个性化实体数据推荐等研究内容，突破了复杂关联场景下传感设备搜索协同技术，解决了“一对多”复杂关联场景下传感设备间的有序协同问题；创新了差异化状态时变实体的观测内容缓存模式，解决了传统无差别缓存模式效率较低的问题；改进了观测内容预测的实体搜索匹配机制，提高了基于观测内容的实体匹配机制的精度。依托课题共发表论文共计13篇（SCI 检索9 篇、EI 检索4 篇）；申请/授权发明专利9件；参加国际学术会议2次；与国内外专家与行业专家交流10次；形成面向物联网实体搜索应用的原型系统1套。本课题的研究成果对我国掌握物联网领域的核心技术自主知识产权具有重要意义，为我国物联网产业技术攻关奠定了理论基础，对推动物联网领域的基础理论创新与智能化应用具有促进意义。