In-System Outlier Smoothing for Real-Time Air Quality Monitoring

The growing concerns about the quality of life have also raised the interests in IAQ (Indoor Air Quality), because IAQ is one of the critical factors affecting our health and life. Therefore, in recent, small size, low cost IAQ monitoring systems are being developed. In particular, a small IAQ monitoring system is based on low cost sensor with lower performance, so that improving the reliability of sensors and coping well with outliers are one of the critical challenges in IAQ monitoring system design. In addition, the existing smoothing algorithms cannot be directly applied to the tiny IAQ monitoring system, since they require huge overhead to be processed in the system. Therefore, in this paper through the experiment with our tiny IAQ monitoring system we identify several outlier problems, and we develop a light-weight in-system smoothing algorithm based on exponential moving average algorithm. The experimental results show that the proposed in-system outlier smoothing algorithm is capable of smoothing outliers well with high accuracy, reliability, and lower memory usage. In addition, through evaluations, we also proved that the EMA based in-system smoothing algorithm outperforms SMA-based method.