Enhanced Air Quality Inference via Multi-View Learning With Mobile Sensing Memory

• Published in: IEEE access Vol. 10; pp. 36616 - 36628
• Main Authors: Liu, Ning, Liu, Xinyu, Lin, Po-Ting, Wang, Yue, Zhang, Lin
• Format: Journal Article
• Language: English
• Published: Piscataway IEEE 2022; The Institute of Electrical and Electronics Engineers, Inc. (IEEE)
• Subjects: Air monitoring; Air pollution; Air quality; Atmospheric modeling; Coders; Encoders-Decoders; Hybrid systems; Learning; Low cost; mobile computing; Monitoring; multi-view learning; Pollution; Sensors; Spatial dependencies; Spatial distribution; spatiotemporal memory; Trajectory; Urban areas
• ISSN: 2169-3536; 2169-3536
• DOI: 10.1109/ACCESS.2022.3164506

Fine-grained air quality can provide essential urban environmental information for administrators and residents. With advances in communication and sensing technologies, low-cost portable sensors installed on vehicles enable high-coverage air quality monitoring. However, data collected by low-cost mobile sensors may be inaccurate and inconsistent in complex operation environments, which brings the issue of data uncertainty. Moreover, due to uncontrolled vehicles and human activities, the coverage of mobile nodes is dynamic over time, leading to uneven or sparse spatial distribution. To address these challenges, we propose AQI-M 3 , a novel framework for fine-grained a ir q uality i nference via m ulti-view learning with m obile sensing m emory. Specifically, an encoder-decoder structure is applied in the region view for modeling the spatial dependencies in pollution maps. More importantly, sensing gradients are extracted in the trajectory view to enable the utilization of uncertain mobile sensing data. In addition, a memory network is designed to capture the spatial patterns from the historical sensing data and provide the global patterns as a complemental guide to overcome dynamic coverage sampling. Extensive experiments are conducted on three real-world deployments of hybrid sensing systems with both static and mobile sensors. Experimental results show that our proposed approach outperforms competitive baselines with 17%<inline-formula> <tex-math notation="LaTeX">\sim 29 </tex-math></inline-formula>% reduction in mean absolute error. Furthermore, detailed evaluations demonstrate the effectiveness and robustness of the proposed framework under dynamic coverage.