基于ASP的智能空间中上下文感知问题的研究

发布时间：2018-05-28 17:43

本文选题：回答集程序 + 智能空间　；参考：《北京工业大学》2016年硕士论文

【摘要】：信息技术的快速发展使计算机软件的执行效率和硬件的性能都得到了很大的提升。在当今时代,很多设备都具有计算能力与数字通信能力,而且这些设备之间可以互相交换信息和使用对方提供的功能。智能空间作为一个融合了计算、信息设备和多模态传感器的工作空间,能够实现设备间的自发交互,但其中的设备、运营商和产品领域种类多样,实现自发交互需要各个参与者之间建立一个共同的标准。因此,相关研究学者研发了Smart-M3平台用于实现智能空间中的自发交互操作,完成设备间信息和服务的共享与存取。在此基础上,Vesa Luukkala等学者为了提高空间的推理能力,将回答集程序ASP与Smart-M3平台进行了整合,采用固定优先关系的方法解决了智能空间中的服务决策问题。但是没有考虑用户上下文信息与环境信息两个影响因素,导致在服务推荐时往往不能满足用户的真实需求。针对这种问题,本文对智能空间中的上下文感知问题进行了比较深入的研究,开展了以下两个方面的工作:(1)针对现阶段采用固定优先关系的方法解决空间中服务决策问题的局限性,本文提出了一种基于回答集程序的智能空间中上下文感知框架,旨在提高空间的动态推理能力。该框架首先建立了一种通用的上下文本体模型,并设计了上下文推理结构;然后定义了上下文推理规则,在规则中利用缺省决策理论动态决策上下文服务的优先关系,与空间中的上下文信息一起组成回答集程序,该程序的回答集对应的是当前上下文信息动态推理的结果。最后,通过引入一个应用实例说明了该框架在具体场景中能有效地完成上下文动态推理,实现了智能空间中具有上下文感知的动态服务决策。(2)进一步的,由于上下文推理过程中使用的原始信息大多来源于空间的物理设备,不可避免的存在信息更新不及时或信息丢失等问题,导致上下文信息出现不一致的情况,从而使推理过程无法进行。目前大多数的上下文不一致检测方法存在着建模复杂的缺点。为了避免这种问题,本文提出了基于元程序的方法检测上下文信息中存在的不一致。首先,使用元程序构建上下文不一致的检测程序,然后使用回答集求解器Smodels求解该程序的回答集,实现上下文信息不一致的自动检测;最后,对检测到的不一致信息执行消除操作策略。这样便解决了上下文信息的不一致性,从而保证上下文推理算法的正确执行。
[Abstract]:With the rapid development of information technology, the execution efficiency and hardware performance of computer software have been greatly improved. In modern times, many devices have computing power and digital communication ability, and these devices can exchange information and use the function provided by the other side. As a workspace that combines computing, information equipment and multimodal sensors, intelligent space can realize spontaneous interaction between devices, but the types of devices, operators and products are diverse. The realization of spontaneous interaction requires the establishment of a common standard among the participants. Therefore, related researchers have developed a Smart-M3 platform to realize spontaneous interaction in intelligent space, and to share and access information and services between devices. On this basis, in order to improve the spatial reasoning ability of Vesa Luukkala and other scholars, the answer set program ASP is integrated with the Smart-M3 platform, and the service decision problem in the intelligent space is solved by the method of fixed priority relation. However, the user context information and environment information are not taken into account, which leads to the failure to meet the real needs of users in service recommendation. In order to solve this problem, the context-aware problem in intelligent space is deeply studied in this paper. This paper presents a context-aware framework in intelligent space based on answer set program, aiming at the limitation of using fixed priority relation to solve the problem of service decision in space. The aim is to improve the spatial dynamic reasoning ability. The framework first establishes a general context ontology model and designs a contextual reasoning structure, then defines the contextual reasoning rules and uses default decision theory in the rules to dynamically determine the precedence of context services. The answer set is composed with the context information in the space. The answer set of the program corresponds to the result of the dynamic inference of the current context information. Finally, an application example is introduced to illustrate that the framework can effectively accomplish contextual dynamic reasoning in specific scenarios, and realize dynamic service decision with context-aware in intelligent space. Because most of the original information used in the process of contextual reasoning comes from the physical equipment of the space, there are inevitable problems such as the information updating is not timely or the information is lost, which leads to the inconsistency of the context information. As a result, the reasoning process can not be carried out. At present, most of the context inconsistent detection methods have the disadvantage of complex modeling. In order to avoid this problem, this paper proposes a meta-program-based approach to detect inconsistencies in context information. First, the meta program is used to construct the context inconsistent detection program, and then the answer set solver Smodels is used to solve the answer set of the program to realize the automatic detection of the inconsistency of the context information. Perform an action elimination strategy for detected inconsistencies. In this way, the inconsistency of context information is solved, and the correct execution of contextual reasoning algorithm is ensured.
【学位授予单位】：北京工业大学
【学位级别】：硕士
【学位授予年份】：2016
【分类号】：TP391.1

【相似文献】