基于人工智能的节能控制物联网云平台的设计与实现

Design and implementation of artificial intelligence-based energy-saving control system on IoT cloud platform

为了解决目前节能控制系统由于设备的有线连接导致的施工困难、改造成本提高和构建系统数据中心导致额外能耗增加等问题,提出了一种采用物联网架构接入云端数据中心的节能控制系统的新方法。该系统由设备层、网络层、平台层以及应用层四层结构构建而成,采用无线低功耗蓝牙技术设计了人体感应传感器、温湿光传感器和电源监控器;采用MQTT协议进行物联网数据的接入;采用机器学习中的强化学习来制定照明控制和空调系统控制的策略。在保证人体舒适度的前提下,达到最大的节能效率。通过网关的边缘计算实现了控制策略的本地运行,使得系统在断网的情况下依然能进行工作。本系统目前已在4所高校投入使用,经测试可实现15%~20%的节能率。

In order to solve the problems of the current energy-saving control system,such as construction difficulties and increasing reconstruction cost caused by the wired connection of the equipment,as well as the increase of additional energy consumption due to the construction of data centers,a new energy-saving control system using the Internet of Things architecture based on the cloud data center is proposed.The system consists of four layers:device layer,network layer,platform layer and application layer.The occupancy motion sensor and temperature-humidity-light sensor are employed in wireless low-power Bluetooth technology.The Internet of Things data is implemented by MQTT protocol.The reinforcement learning,one of machine learning methods is used to develop policies for lighting and air conditioning control.Under the requirement of ensuring comfortability,the maximum energy-saving efficiency is achieved.The local operation of the control policy is implemented through the edge computing of the gateway,so that the system can still work without network.This system has been put into use in four universities,and the test reveals that it can achieve 15%?20%energy saving rate.