A symbolic programming approach to intelligent data reduction

An intelligent filtering system for filtering notifications from sensors on the basis of the user's current focus and the relative importance of the notification has been developed to reduce the telemetry data from the Hubble Space Telescope (HST) Electrical Power System testbed. The motivation, design, and development of this system are discussed. A particular area of interest is the health and performance of the six nickel-cadmium batteries as they undergo the charge and load functions of the simulated orbiting. The knowledge base in the present system is derived primarily from the warnings and alarms in NICBES, an expert system for health management and diagnosis of these batteries. Both systems get their input from a stream of 370 sensor readings emitted from the testbed every minute. Included in these streams are cell voltages and pressures, total battery voltages and currents, and bus voltages and currents.<
>