改进型卡尔曼滤波的PM2.5浓度预测系统  

PM2.5 concentration prediction system based on improved Kalman filtering

作　　者：许江淳[1] 余丽玲[1] 岳秋燕 任杰[1] 

机构地区：[1]昆明理工大学信息工程与自动化学院,云南昆明650500

出　　处：《传感器与微系统》2017年第7期106-109,共4页Transducer and Microsystem Technologies

基　　金：国家质检总局科技计划资助项目(2013QK104)

摘　　要：针对目前室内空气净化器功能单一、自动调节能力弱的特性,设计了一种改进型卡尔曼滤波对PM 2.5浓度进行预测的系统,为净化器提供可靠的调节参数。选用S3C2440芯片,通过外设采集PM2.5浓度和温度,以及对历史数据进行分析。通过改进型卡尔曼滤波对历史数据、PM2.5浓度以及温度进行融合,可以准确得到下一时刻的浓度值。测试结果表明:系统运行稳定,算法跟踪效果明显,鲁棒性强,可以准确预测出下一时刻的PM2.5浓度,并且预测误差在2.0%以内,满足设计要求。In view of the characteristics of current indoor air purifier function is single and the automatic regulation ability is weak, an improved Kalman filtering system is designed to predict the concentration of PM2.5, which provides a reliable regulation parameter for air purifier. The S3C2440 chip is used to collect PM2. 5 concentration and temperature, and analyze on historical data. On the basis of the system, through the improved Kalman filter to fuse historical data, PM2. 5 concentration and temperature can be accurately obtained concentration at the next moment. The results of test show that the system runs stably with its effective algorithm, strong robustness, and it can accurately predict the PM2.5 concentration at the next time point. The prediction error is less than 2.0 % , which can meet the requirements of design.

关 键 词：卡尔曼滤波 PM2.5 净化器 S3C2440芯片 浓度值 鲁棒性 

分 类 号：TP311[自动化与计算机技术—计算机软件与理论]