基于SVM的智能家居辅助决策系统研究

发布时间：2018-05-07 03:44

本文选题：智能家居 + 辅助决策　；参考：《华北电力大学》2017年硕士论文

【摘要】：随着生活水平的提高及物联网、机器学习等技术的发展,如何使家居具备自主“学习”能力,根据用户个人习惯自主控制照明灯、窗帘等家居的开闭成为当前智能家居研究的热点之一,对智能家居的发展及推广具有重要意义。本文对智能家居辅助决策中的异常数据处理、获取用户决策规则集、用户类别判定及优化决策等问题进行了深入的研究。本文主要工作如下:1)分析和研究了智能家居辅助决策中的环境信息特点、决策方法、评估标准等,给出了一种支持向量机和K近邻算法相结合的智能家居辅助决策模型。该模型构建主要包括个人环境信息建模、异常数据处理、相似用户查找及基于支持向量机的用户决策规则集获取等。通过实验,与一般决策系统、基于BP神经网络的决策系统进行决策准确度的对比,结果表明,该方法具有耗时少、精度高的特点。2)给出了一种基于自适应步长和增强随机扰动的的改进萤火虫算法,用于优化决策系统,通过引入亮度值和欧几里得距离公式改进了原算法中的位置更新公式,并利用matlab中的测试函数对该算法进行了测试,实验证明该算法在避免收敛早熟及无法收敛问题上有一定的优势。3)设计并实现了该智能家居辅助决策原型系统,研究内容主要包括智能家居系统总体架构设计、个人环境信息的采集及用于查看和控制家居状态的智能家居APP的设计与开发。
[Abstract]:With the improvement of living standard and the development of technology such as Internet of things and machine learning, how to make the home have the ability of autonomous "learning" and control the lighting independently according to the user's habits. Curtain and other household opening and closing has become one of the hot spots in the research of smart home, which is of great significance to the development and promotion of smart home. In this paper, the problems of abnormal data processing, obtaining user decision rule set, user category decision and optimization decision in intelligent home assistant decision are deeply studied. The main work of this paper is as follows: (1) this paper analyzes and studies the characteristics of environmental information, decision methods, evaluation criteria and so on in the intelligent home assistant decision making, and gives a kind of intelligent home assistant decision model combining support vector machine and K-nearest neighbor algorithm. This model mainly includes personal environment information modeling, exception data processing, similar user lookup and user decision rule set acquisition based on support vector machine. Compared with the general decision system and the decision system based on BP neural network, the experimental results show that the method has less time consuming. (2) an improved firefly algorithm based on adaptive step size and enhanced random perturbation is presented to optimize the decision-making system. By introducing the luminance value and Euclidean distance formula, the position updating formula in the original algorithm is improved. The test function in matlab is used to test the algorithm. The experiment shows that the algorithm has some advantages in avoiding premature convergence and unconvergent problem. 3) the intelligent home aided decision making prototype system is designed and implemented. The main contents of the research include the overall architecture design of smart home system, the collection of personal environment information and the design and development of intelligent home APP for viewing and controlling the status of home.
【学位授予单位】：华北电力大学
【学位级别】：硕士
【学位授予年份】：2017
【分类号】：TU855

【参考文献】