面向情景感知的智能物体语义建模与自适应控制研究

How the smart objects to adapt to the dynamic changes in environment for these smart objects and IoT applications adaptively controlling the context-aware environment is a challenge for mobile distributed computing and context-aware computing. In allusion to the requirement of unified identification, sematic abstraction, context-aware modeling and self-adaptive controlling for the smart objects, the project studies the unified identification mechanism and resource description model of the IoT resource objects, builds semantic abstraction and defines mapping relations between native devices and heterogeneous services, and proposes a dynamic service registration and discovery mechanism for the smart objects in context-aware environment. This project establishes context ontology library, user and smart objects ontology for context modeling of the smart object applications and end user services. To select and invoke the matched service set dynamically, we study user information and scenario information related mining algorithms for automatic matching the pre-defied context models and control rules. With the help of ontology rule reasoning engine based on RETE algorithm, the self-adaptive control of the smart objects can be realized. Meanwhile, taking privacy protection of IoT data and access control into consideration, an M2M self-adaptive control system architecture and prototype system will be proposed in this project.

智能物体如何适应周围环境的动态变化，以便让智能物体及其物联网应用能够自适应控制情景感知环境，这对移动分布式计算和情境感知计算是一个挑战。针对智能物体的标识、语义抽象、情景感知建模与自适应控制需求，项目研究物联网资源标识体系与资源描述模型，对物联网原生设备和异构服务进行统一语义抽象和映射关系定义，提出感知情景中智能物体的动态服务注册与发现机制。研究物联网环境下的情景本体库、设备本体、用户本体及用户行为模型的构建方法，对智能物体应用情景模型和终端用户群的服务模型进行情景建模，研究用户和情景信息关联挖掘算法进行自动匹配预先定义的情景模型和控制规则，基于本体规则的RETE算法推理出情景匹配的设备控制服务，实现智能物体的环境自适应控制。研究物联网数据隐私保护与访问控制机制，提出智能物体自适应控制体系架构并研发原型系统。