Assessing data veracity for data-rich manufacturing

Data-driven decision making is becoming more important in today's data-rich manufacturing environments. Just as the quality of the material is important to guarantee a quality final product, quality of data is crucial to make correct and efficient decisions. Data veracity refers to the trustworthiness of data. In a modern manufacturing facility, data is often produced by machine sensors. Due to sensor degradations, failures or possible hacking attacks, data veracity may be compromised that causes shop floor decision making to be inaccurate. In addition, data may be incomplete to cause uncertainty in data analytics. Many researchers have studied different aspects and measurement of data veracity including timeliness, completeness, accuracy and consistency. But none of the current methods provide trustworthiness assessment on data generated over time. This paper introduces a model for fusing multiple data sources based on a multi attribute decision making (MADM) method with the consideration of uncertainty. A numerical example is provided to demonstrate the use of the proposed method.