基于改进卡尔曼算法的室内温度数据融合  

Indoor Temperature Data Fusion Based On Improved Kalman Algorithm

作　　者：刘陈男 罗恒[1,2] LIU Chennan;LUO Heng(School of Electronics and Information Engineering,Suzhou University of Science and Technology,Suzhou 215000,China;Jiangsu Key Laboratory of Building Intelligence and Energy Conservation,Suzhou 215000,China)

机构地区：[1]苏州科技大学电子与信息工程学院,江苏苏州215000 [2]江苏省建筑智慧节能重点实验室,江苏苏州215000

出　　处：《计算机测量与控制》2024年第1期179-184,共6页Computer Measurement &Control

基　　金：国家自然科学基金项目(61602334);国家自然科学基金项目(51874205)。

摘　　要：针对传统建筑物内部空间结构复杂,布线成本高,数据采集精度不高等问题,采用LoRa无线通信技术构建传感器网络,主要用于监测室内温度参数的变化;对于传统的卡尔曼数据融合结果存在较小波动的现象,引入了孤立森林算法,提出了基于改进的卡尔曼滤波算法的室内温度数据融合算法;通过在采集到的数据集中随机添加扰动样本和畸变数据,对3种算法产生的误差进行比较,改进的卡尔曼数据融合算法在有扰动样本的情况下,误差范围控制在-0.12~0.1之间,在带有畸变数据时,误差范围在-0.03至0.14之间,均远小于传统的卡尔曼数据融合算法和平均值算法;实验仿真的结果表明,改进的算法提高了室内温度数据采集的鲁棒性和准确性。For the problems of complex internal space structure of traditional buildings,high wiring cost and low data acquisition accuracy,LoRa wireless communication technology is used to build a sensor network,which is mainly used to monitor the changes of indoor temperature parameters;For the phenomenon of small fluctuations in traditional Kalman data fusion results,the isolated forest algorithm is introduced,an indoor temperature data fusion algorithm based on the improved Kalman filter algorithm is proposed.The collected data is randomly added to the perturbed samples and distorted data,the generation errors of three algorithms are compared,The error range of the improved Kalman data fusion algorithm is controlled between-0.12 and 0.1 with the perturbed samples and from-0.03 to 0.14 with the distorted data,the improved Kalman data fusion algorithm is much smaller than the traditional Kalman data fusion algorithm and average algorithm.The experimental simulation results show that the improved algorithm increases the robustness and accuracy of indoor temperature data acquisition.

关 键 词：LoRa技术 孤立森林算法 数据融合 卡尔曼滤波算法 无线传感器网络 

分 类 号：TP212[自动化与计算机技术—检测技术与自动化装置]