WCSE

Social Activities Assist Sensor Network through Multi-modal Information Fusion

Abstract— In China, environment issue has been attached greater importance, especially the management and protection in air quality. There are already sensor networks existing and locating in the major cities for PM2.5 monitoring. These sensor networks provide the real-time data of PM2.5 at different locations which are further used to publish and predict the future air quality. In this study, we propose a new frame work to fuse the multi-modal data sources, i.e. data stream from PM2.5 sensor networks and the real-time social activities from WeChat platform, for air quality assessment in real-time. Both semantic analysis and 2D interpolation methods are used in the multi-modal fusion procedure. As an application, we construct the cloud data server based on SinaAppEngine (SAE) and develop the service system using PHP and MySQL. Based on the new system, social activities were automatically transferred into quantitative evaluations of local air quality. Then using the GPS information of social activities, the evaluated air quality was fused with the sensor data and further interpolated on the map. The accuracy and robustness of our system were tested based on the data in the university campus. A one-month continuous monitoring was also performed and the detailed analysis demonstrated the monitoring performance of the new system. The tools and methods proposed in this study show more potentials in environmental monitoring, policy decision making and sociological research.

Index Terms— social activities, sensor network, data fusion, PM2.5

Jia Wu, Zhe Yu, Zhaocheng Su, Lingke Zhang, Miao Li, Peng Miao
School of Communication and Information Engineering, Shanghai University, CHINA
Zhe Yu, Lingke Zhang, Peng Miao
Institute of Biomedical Engineering, Shanghai University, CHINA

[Download]

Cite: Jia Wu, Zhe Yu, Zhaocheng Su, Lingke Zhang, Miao Li, Peng Miao, "Social Activities Assist Sensor Network through Multi-modal Information Fusion," Proceedings of 2017 the 7th International Workshop on Computer Science and Engineering, pp. 584-589, Beijing, 25-27 June, 2017.