Trustworthy data collection for cyber systems: A taxonomy and future directions

Abstract

Due to technology limitation and environmental influence (i.e., equipment faults, noises, clutter, interferences, and security attacks), the sensor data collected by Cyber-Physical System (CPS) is inherently noisy and may trigger many false alarms. These false or misleading data can lead to wrong decisions. Therefore, data trustworthiness (i.e., the data is free from error, up to date, and originate from a reputable source) is always preferred. However, it often has high cost and challenges to identify fault, noise, cyber-attack, and real-world facts, especially in heterogeneous and complex IoT environment. In this article, we briefly review the current developments and research trend in this research area. We highlighted all the challenges and potential solutions for the trustworthy data collections in CPS and propose a taxonomy for data trustworthiness in CPS. Taxonomy aims to describe different aspects of research in this field. Furthermore, it will help researchers as a reference point for the design of data reliability and data trustworthiness evaluation methods. Based on the observations, future directions are also suggested.