DISCRIMINATIVE HIDDEN KALMAN FILTERS FOR CLASSIFICATION OF STREAMING SENSOR DATA IN CONDITION MONITORING

摘要

A method for monitoring a condition of a system or process includes acquiring sensor data from a plurality of sensors disposed within the system (S41 and S44). The acquired sensor data is streamed in real-time to a computer system (S42 and S44). A discriminative framework is applied to the streaming sensor data using the computer system (S43 and S45). The discriminative framework provides a probability value representing a probability that the sensor data is indicative of an anomaly within the system. The discriminative framework is an integration of a Kalman filter with a logistical function (S41).

主权项

1. A method for monitoring a condition of a system or process, comprising:
acquiring sensor data from a plurality of sensors disposed within the system; streaming the acquired sensor data in real-time to a computer system; and applying a discriminative framework on the streaming sensor data using the computer system, the discriminative framework providing a probability value representing a probability that the sensor data is indicative of an anomaly within the system, wherein the discriminative framework is an integration of a Kalman filter with a logistical function.