Intelligent real-time fault diagnosis of greenhouse sensors

Abstract An intelligent real-time system has been developed to monitor greenhouse sensor signals. Its objective is to detect and correct errors in the data due to faulty sensor readings. The system (FLASH- Flexible Artificial Intelligence for Signal Handling in real time) uses a novel knowledge-based approach for fault detection and isolation. It comprises three modules: Scheduler, Signal Processing and Reasoning, which communicate via a Blackboard. FLASH has been implemented in C++ on a Sun Workstation running Solaris 2, and it was demonstrated using real-time data.