Refine
Year of publication
- 2016 (2) (remove)
Document Type
- Conference Proceeding (2) (remove)
Has Fulltext
- yes (2)
Is part of the Bibliography
- yes (2)
Keywords
Institute
- Fakultät IuI (1)
The wide distribution of smart phones allows to inform and interact with citizens in real-time, thus enabling the vision of smart cities. However, the reliability of smart city applications highly depends on the availability of appropriate, accurate, and trustworthy data. To increase the reliability of smart city applications, the European project CityPulse employs knowledge-based methods for monitoring and testing at all stages of the data stream processing and interpretation pipeline. During design-time testing validates the behaviour of applications with regard to different levels of quality of information. During run-time monitoring assesses the reliability of data streams, the plausibility of information, and the correct evaluation of extracted events. The monitored quality is exploited by fault recovery and conflict resolution mechanisms to ensure fault-tolerant execution of applications.
Reliable information processing is an indispensable task in Smart City environments. Heterogeneous sensor infrastructures of individual information providers and data portal vendors tend to offer a hardly revisable information quality. This paper proposes a correlation model-based monitoring approach to evaluate the plausibility of smart city data sources. The model is based on spatial, temporal, and domain dependent correlations between individual data sources. A set of freely available datasets is used to evaluate the monitoring component and show the challenges of different spatial and temporal resolutions.