An Approach for Assessing Industrial Iot Data Sources to Determine Their Data Trustworthiness

Trustworthy data in the Industrial Internet of Things are paramount to ensure correct strategic decision-making and accurate actions on the shop floor. However, the enormous amount of industrial data generated by a variety of sources (e.g. machines and sensors) is often of poor quality (e.g. unreliable sensor readings). Research suggests that certain characteristics of data sources (e.g. battery-powered power supply and wireless communication) contribute to this poor data quality. Nonetheless, to date, much of the research on data trustworthiness has only focused on data values to determine trustworthiness. Consequently, we propose to pay more attention to the characteristics of data sources in the context of data trustworthiness. Thus, this article presents an approach for assessing Industrial Internet of Things data sources to determine their data trustworthiness. The approach is based on a meta-model decomposing data sources into data stores (e.g. databases) and providers (e.g. sensors). Furthermore, the approach provides a quality model comprising quality-related characteristics of data stores to determine their data trustworthiness. Moreover, a catalogue containing properties of data providers is presented to infer the trustworthiness of their provided data. An industrial case study revealed a moderate correlation between the data source assessments of the proposed approach and experts.