基于智能空间的服务方法与技术应用研究

发布时间：2018-01-09 12:30

本文关键词：基于智能空间的服务方法与技术应用研究　出处：《北京科技大学》2017年博士论文　论文类型：学位论文

【摘要】：智能空间是一个集用户、环境、设备和服务为一体的人机交互系统。目前,智能空间的实现受以下几个方面的制约:异构设备存在交互壁垒,智能设备资源受限;智能空间的服务模式缺乏主动性和智能性;智能空间服务中缺乏有效的资源管理和利用手段。本文使用REST风格的微服务技术,融入基于状态机的设备驱动技术,结合基于XMPP的信息推送技术,设计了智能空间服务框架。在此基础上,本文主要的研究内容和创新点如下:(1)提出了一种基于模糊粗糙集的智能服务规则静默提取与执行方法,以实现具有用户个性化喜好的产生式规则推理服务。首先,通过表驱动的状态机,把历史用户数据用状态表表示;其次,使用模糊粗糙集理论对历史数据表进行约简,把约简结果作为用户服务规则;再次,引入二进制蝙蝠算法,用于启发式规则约简过程。通过改进蝙蝠位置更新策略,改进了算法性能;最后,使用并改进了基于Rete算法的规则引擎,激活产生式规则推理过程,实现和谐的用户个性化服务。实验结果表明,本文方法能够在无需用户主动干预的情况下,快速准确提取用户个性化服务规则,提升了智能空间服务的特异性和舒适性。(2)提出了一种基于云模型的用户位置追踪节能算法,以有效减少智能空间服务过程的能源消耗。首先,针对无线传感器网络定位产生的能源开销过大问题,提出预测用户移动轨迹,智能地激活-休眠相关传感器的方法,来减少传感器能耗;其次,提出使用云模型描述用户的移动特点,来预测用户位移序列;再次,针对计算量随时间增加和计算资源有限这一矛盾,提出增量式迭代计算方法,减少计算量;最后,利用质心定位方法计算用户位置。实验结果表明,本文算法有效减少了定位服务的能源消耗,提高了系统的定位追踪性能,实现了智能服务感知的绿色节能。(3)提出了一种基于改进人工蜂群算法的具有QoS保证的复杂服务快速选择算法,以有效改善智能空间组合服务的搜索速度和精度。首先,将服务选择问题抽象为整数规划问题,使用服务QoS指标计算用户满意度,并加入惩罚因子,将多目标优化问题转化为单目标优化;其次,使用人工蜂群算法解决组合服务寻优过程的"组合爆炸"问题;再次,提出了交叉选择因子,并通讨论了其设计策略,从而提高算法收敛速度和收敛精度;最后,分析了算法收敛性,给出了算法步骤。实验结果表明,本文算法能够有效提高服务组合的搜索速度和精度,提高用户对组合服务的主观满意度。最后,将本文研究的面向智能空间的服务技术与方法,融入到智能家居环境与智能养老社区环境中,进行相关示例应用。实验结果表明,本课题的研究成果为智慧家庭、智慧社区走入人们的生活做出了贡献,为智能空间的研究提供了有效的支撑平台。
[Abstract]:Smart space is a set of users, environment, man-machine interaction system of equipment and services as a whole. At present, intelligent space is restricted by the following aspects: heterogeneous devices exist interaction barriers, smart devices with limited resources; smart space service mode lack of initiative and intelligence; intelligent space in lack of effective resource management service and by means of using the REST technology. The micro service style, into the driving technology of state machine equipment based on combination of XMPP information push technology based on the design of intelligent space service framework. On this basis, the main research content and innovation are as follows: (1) this paper proposes a fuzzy rough set intelligent service rules extraction method based on silence and execution, to achieve a user preference generative rule reasoning service. First of all, through the state machine table driven, the history of user data Represented by the state table; secondly, using fuzzy rough set theory on historical data table reduction, the reduction of the user service rules; thirdly, introducing the binary bat algorithm, heuristic rules for reduction process. By improving the bat location update strategy, which improves the performance of the algorithm; finally, using the Rete algorithm and rule engine based on improved the activation of generative rules, the reasoning process, realize the personalized service of harmony. The experimental results show that this method can active without user intervention, quickly and accurately extract the personalized service rules of users, enhance the intelligent space service specific and comfort. (2) proposed a cloud user location algorithm the tracking model based on intelligent space services to effectively reduce the energy consumption of the process. Firstly, based on the positioning of wireless sensor network energy cost is too large to ask This problem, predict the user movement trajectory, intelligent activation method of dormancy related sensors, to reduce the energy consumption of sensor; secondly, the characteristics of mobile use cloud model to describe the user, users to predict the displacement sequence; thirdly, according to the amount of computation increases with time and limited computing resources of this contradiction, proposes the calculation method of incremental iteration, reduce the amount of computation; finally, calculate the position of the user using the centroid location method. The experimental results show that this algorithm can effectively reduce the energy consumption of the positioning service, improve the system tracking performance, realizes the green energy-saving intelligent service perception. (3) proposed an improved artificial bee colony algorithm has fast complex QoS service guarantee the selection algorithm based on intelligent space to improve the search speed and accuracy of composite service. First, the service selection problem into an integer programming problem, the use of Q services The calculation of user satisfaction index oS, and add the penalty factor, the multi-objective optimization problem is transformed into a single objective optimization; secondly, the use of artificial bee colony algorithm to solve the service composition optimization process the combinatorial explosion problem; again, put forward the cross selection factor, and discusses its design strategy, so as to improve the convergence speed and convergence finally, the analysis accuracy; the convergence of the algorithm, the algorithm steps are given. The experimental results show that this algorithm can effectively improve the service combination of search speed and accuracy, improve the user of the composite service subjective satisfaction. Finally, the technology and method of service oriented smart space research, integrated into the smart environment and Home Furnishing intelligent pension community environment, for example application. The experimental results show that the research results for the smart home, smart community into people's life to make a contribution to wisdom The research of energy space provides an effective support platform.

【学位授予单位】：北京科技大学
【学位级别】：博士
【学位授予年份】：2017
【分类号】：TP18;TP11

【参考文献】