A correlation-based approach to trustworthy sensing for cyber-physical systems

Trustworthiness of sensing is of great importance to the monitoring and control of Cyber-Physical Systems (CPS), because it provides the physical foundation for ascertaining the performance of the system being monitored, based on which control actions are executed. Since sensors in CPS are often placed in different locations, current methods that depend on redundant information from sensors at the same location for trustworthiness evaluation is not suitable for such applications. A novel correlation-based approach is proposed in this paper to determine the trustworthiness of sensing. Specifically, the approach first exploits correlations between sensors by mapping all data to a base sensor, which can be chosen randomly. Then, trustworthiness of each sensor is determined using integrated information of these mapped data. Two case studies are investigated to verify the effectiveness and robustness of the proposed approach for trustworthy sensing.