智能空间下基于本体技术的机器人服务自主决策及执行

发布时间：2018-02-11 03:25

本文关键词： 服务机器人智能空间本体技术服务认知任务规划　出处：《山东大学》2017年硕士论文　论文类型：学位论文

【摘要】：服务机器人逐渐融入人类的日常生活,在家庭、医疗、办公等领域扮演着越来越重要的角色,因此人们对机器人的服务能力提出了更高的要求,希望机器人能够主动地为人类提供更加智能、个性化的服务。机器人对知识的获取和利用能力是机器人实现正确的场景感知、理解用户需求从而主动提供智能化服务的保障。智能空间技术使机器人的感知能力和信息处理能力得到了极大的提升,但是空间资源的多样性、异构性和动态性阻碍了知识的融合和共享,使机器人难以实现对空间知识的充分利用,因此如何对知识进行有效的组织,提高智能空间知识的表示能力,为机器人的服务决策和任务执行提供有力的支持成为一个重要的研究内容。本文在智能空间环境下,设计了基于智能空间本体的机器人服务自主决策及规划执行方法,使机器人在无需人为干预的条件下,主动地为用户提供个性化的服务,主要工作如下:1、研究本体建模技术,根据智能空间的资源及机器人的服务能力,为智能空间划分本体领域:环境领域、用户领域和服务领域,将空间中的对象信息在各领域中进行匹配,得到智能空间本体。利用关系数据库作为智能空间本体的存储工具,形成智能空间本体库,并在本体库的基础上,实现知识的动态更新。2、考虑到智能空间本体库中知识的局限性,利用Jena对本体库中的知识进行扩展。通过建立自定义规则的方式,挖掘空间中隐藏的逻辑知识;利用模糊集的思想,建立"数据-概念"转换机制,将本体库中的数据信息映射为能够反映用户喜好或者服务质量的概念表示。扩展后的智能空间本体库,能够为机器人的服务推理提供更完备的场景信息。3、分析机器人服务的特征和决定性因素,采用SWRL本体规则描述语言建立服务推理规则库,通过事实与规则相匹配的方法,实现机器人服务的自主推理、决策。结合服务推理结果和智能空间本体库中的空间状态知识,生成JSHOP2规划器的问题描述,利用分层任务网络规划算法,完成服务任务的规划,为机器人提供需要执行的动作列表。4、将本文所提出的方法,加载到山东大学服务机器人实验室自主研发的服务机器人上,在智能空间场景下进行实验测试,对方法的有效性和实时性进行验证。
[Abstract]:Service robots are gradually integrated into human daily life and play more and more important roles in the fields of family, medical treatment, office and so on. Therefore, people put forward higher demands on the service ability of robots. It is hoped that robots can actively provide more intelligent and personalized services for human beings. The ability of robots to acquire and utilize knowledge is to realize the correct perception of the scene. Understanding the needs of users in order to actively provide intelligent services. Intelligent space technology has greatly improved the perception and information processing ability of robots, but the diversity of space resources, Heterogeneity and dynamics hinder the integration and sharing of knowledge, which makes it difficult for robots to make full use of spatial knowledge. Therefore, how to organize knowledge effectively and improve the representation ability of intelligent spatial knowledge. It has become an important research content to provide powerful support for robot service decision and task execution. In the intelligent space environment, an autonomous decision making and planning method for robot service based on intelligent space ontology is designed in this paper. Under the condition of no human intervention, the robot can provide personalized service to the user actively. The main work is as follows: 1. Research ontology modeling technology, according to the resources of intelligent space and the service ability of robot. The domain of ontology is divided into three parts: environment domain, user domain and service domain. The object information in the space is matched in each domain to get the intelligent space ontology. The relational database is used as the storage tool of the intelligent space ontology. The intelligent space ontology library is formed, and on the basis of the ontology library, the knowledge is dynamically updated. Considering the limitation of the knowledge in the intelligent space ontology library, the knowledge in the ontology library is extended by Jena. Mining the hidden logic knowledge in the space, using the idea of fuzzy set, establishing the conversion mechanism of "data-concept", Mapping the data information in the ontology library to the concept representation that can reflect the user's preferences or the quality of service. It can provide more complete scene information for robot service reasoning, analyze the characteristics and decisive factors of robot service, establish the service inference rule base by SWRL ontology rule description language, and use the method of matching the facts and rules. In order to realize autonomous reasoning and decision making of robot service, the problem description of JSHOP2 planner is generated by combining the result of service reasoning with the knowledge of spatial state in intelligent space ontology library, and the service task planning is accomplished by using hierarchical task network planning algorithm. The method proposed in this paper is loaded into the service robot developed independently by the service robot laboratory of Shandong University, and the experimental test is carried out in the intelligent space scene. The validity and real-time of the method are verified.
【学位授予单位】：山东大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TP242

【相似文献】