基于规则控制与XGBoost的智能开窗系统设计

DESIGN OF INTELLIGENT WINDOWING SYSTEM BASED ON RULE CONTROL AND XGBOOST

为了提高开窗系统在雨天及初始工作时的智能性,降低开窗器的能耗损失,设计一款基于规则控制和XGBoost算法的智能开窗系统。系统通过使用ZigBee技术、使单片机处于停止模式和切断非实时工作模块供电电路降低电路功耗,并综合考虑当前时间、环境和未来1小时的降雨概率,通过融合规则控制和XGBoost算法预测当前窗户状态。实验结果表明,该系统能在降低功耗的同时实现远程控制和防雨水入户等基础功能,且能学习用户的开关窗习惯,在融合使用规则控制和XGBoost算法及综合考虑当前和未来环境后,提高系统的预测准确性。

In order to improve the intelligence of the window opening system in rainy days and during the initial operation of the system,and reduce the energy loss of the window opener,an intelligent windowing system based on rule control and XGBoost is designed.The system reduced the power consumption of the circuit by using ZigBee technology,making the singlechip in stop mode and cutting off the power supply circuit of the nonrealtime working module.The system comprehensively considered the current time,environment and the probability of rain in the next hour,and predicted the current state of the window by combining rule control with XGBoost.The experimental results show that the system can achieve basic functions such as remote control and preventing rainwater from entering the room while reducing power consumption,and can learn the user's habit of opening and closing windows.The prediction accuracy of the system is improved by combining rule control with XGBoost and considering the current and future environment comprehensively.