室内环境的数据融合自适应调控方法研究

针对传统的室内环境控制方式存在抗干扰性差、耗能高、智能化程度低及融合性简单等不足,提出一种通过多数据融合技术实现的新型智能化调控方法.从卡尔曼滤波器的基本原理出发,采用多维线性滤波模型对多个传感器采集的环境指标数据进行滤波,得到当前环境状态的有效估计,然后通过神经网络进行融合,得到当前状态期望值的最优估计,最终通过动态矩阵控制(DMC)算法来根据性能指标完成优化控制.仿真结果表明,相较于现有的室内环境控制方法,此方法的滤波误差<2%,调控误差<0.01%,节能率> 1%,从而能够有效提高控制精度和抗干扰能力,在尽量降低能耗的基础上实现室内环境的最优化控制,可以推广应用于对环境因素要求比较苛刻的场合的精细调节,比如医院、温室、科研实验室等.

Research on adaptive control method for indoor environment base on data fusion

In view of poor anti-interference, high energy consumption, low intelligence and integration of traditional indoor environment control methods, a new intelligent control method based on multi-data fusion technology is proposed.Firstly, according to the basic principle of Kalman filter, a multi-dimensional linear filtering model is established to filter the raw data collected by multiple sensors to obtain a valid estimate of current environment state. Then, the filtered data fused by the neural network to obtain the optimal estimate of expectation value in current state. Finally, the dynamic matrix control(DMC) algorithm is applied to complete the optimal control according to the performance indicators. The simulation results indicate that compared with the existing indoor environment control methods, the filtering error rate of this method is below 2%, the adjustment error rate below 0.01%, and the energy saving rate above 1%. This method can effectively adjust the indoor environmental parameters with low energy consumption, and contributes to improving the control precision and anti-interference ability. Moreover, it can be applied to some environmental sensitive occasions where require fine adjustments, such as hospitals, greenhouses, and laboratories, etc.