基于DAG-SVM算法的城市路灯照明系统的研究
The Study on Lighting System of Urban Street Lamp Based on DAG-SVM Algorithm

针对目前城市路灯照明采用手动、感应式和定时控制等路灯控制方式引起电能浪费及智能化程度低的问题,设计了一个新的城市路灯照明系统.该系统由路灯节点控制器、集中控制器、云服务器和后台管理系统共4层结构组成.依托该系统结构,提出有向无环图支持向量机(DAG-SVM)的6分类调光算法.实验结果表明:与手动、感应式和定时控制方式相比,采用DAG-SVM算法的路灯照明系统不仅调光更智能准确,而且节能效率分别提升57.5%、14.5%和5.0%,且系统节能达到63%.
Aiming at the power waste and low intelligent degree after using the manual, inductive and timing control methods, the lighting system of urban street lamp is studied. The system consists of fourtier structures of street node controller, centralized controller, cloud server and backstage management system. Depended on the system structure, the six classification dimming of directed acyclic graph support vector machine(DAG-SVM)algorithm is proposed. Firstly, the six classification hyperplanes are constructed according to the different environment data around street lamps; secondly, the six classification dimming model by using the classification hyperplane training is used to judge the dimming level of street lamps. Experimental results show that compared with the manual, inductive and timing control methods, the street lighting system adopted DAG-SVM algorithm can not only be more intelligent and accurate, but also improve the system energy efficiency increased by 57.5 percent, 14.5 percent and 5.0 percent, saved up to 63 percent