云计算
计算机科学
家庭自动化
软件部署
语音指挥设备
过程(计算)
互联网
边缘计算
GSM演进的增强数据速率
人工神经网络
控制(管理)
多媒体
实时计算
嵌入式系统
机器学习
人工智能
万维网
电信
操作系统
出处
期刊:International Journal of Emerging Electric Power Systems
[De Gruyter]
日期:2022-10-12
卷期号:23 (6): 873-885
标识
DOI:10.1515/ijeeps-2022-0124
摘要
Abstract With the rapid progress and wide application of cutting-edge technologies such as automatic control technology and wireless communication technology in recent years, smart home, as an effective combination of these cutting-edge technologies and daily life, has received more and more attention and Research has been greatly developed. In order to solve the shortcomings of the existing home control need to download various cumbersome APP, the WeChat applet is introduced into the smart home system, and the intelligent voice technology is planned to be used in the smart home control to realize the voice control, improve the user experience and make the home Life gets smarter. This time, the deep learning technology was used in the development process of the smart home control system, and according to the historical information of the home, combined with the learning ability of the recurrent neural network for time series data, through the mining, analysis and learning of historical data, the construction based on specific The user’s unique computing model can seamlessly connect the capabilities of the voice cloud platform with the capabilities of the Internet of Things cloud platform through the Web server, making it possible to quickly access the voice recognition capabilities to smart homes. The entire system has stable connectivity, easy deployment and low cost. The main functions are deployed in the cloud and extended. After 10 rounds of iterative training of the Attention-GRU model in this paper, its prediction accuracy can quickly rise to about 97%, and finally stabilize at about 98.2%, and the lighting prediction accuracy can reach 95% or higher.
科研通智能强力驱动
Strongly Powered by AbleSci AI