Real-time automated diagnosis for human-computer based monitoring and control systems

Many real-time monitoring and control systems are designed to be computer-based but to still retain significant roles for humans in operating and supervising the performance of the overall system. Examples of this type of monitoring and control system may be found in control rooms for wind tunnels, mission control rooms for spacecraft, and control rooms for power plants and manufacturing plants. These monitoring and control systems may also be found on board research and surveillance aircraft. The requirements of the human operator's supervisory role in these systems may tax the physiological limits of human mental concentration and physical endurance. These human limitations could negatively influence the safety and performance capabilities of the overall system. This paper describes a prototype System Health Management (SHM) system which was designed to automate some of the human monitoring and diagnosis tasks for the operations crew of the Kuiper Airborne Observatory (KAO), an airborne telescope system. The intent in developing such a system is to ease the burden on the human operators by automating the more monotonous tasks that are difficult for humans to perform well over an extended period of time, thereby freeing them to devote their concentration to the more skill-oriented tasks at which they tend to excel. A second intent is to assist the human operators in quickly diagnosing and repairing failures in flight. The combined effect of these benefits is expected to be better efficiency in the operation of the overall system (i.e. fewer human operators needed), and a reduction of lost science (data acquisition) due to equipment failures.