Air Quality Estimation based on Multi-Source Heterogeneous Data from Wireless Sensor Networks

It's a great challenge to offer a fine-grained and accurate air quality monitoring service in urban areas limited to the cost of the professional facilities. With the development of the wireless sensor networks (WSNs), it brings new opportunity to achieve this goal at low cost. However, WSNs are quite different on temporal and spatial distribution, some WSNs even have irregular real-time features, which makes it a hard problem to use the data collected from different WSNs to achieve the same goal. In this paper, we propose a framework for air quality estimation based on multi-source heterogeneous data collected from WSNs. We collect five kinds of data from different sources in real world, including the data with irregular real-time features. We divide the data sets into three sub classifiers to make the analysis and get the final results with an extreme learning machine (ELM) based multilayer perceptron model. The results show that our method outperforms other methods, and the precision of classification can be 90.8%.