DYNAMIC ACROSS TIME AUTONOMOUS - SENSING, INTERPRETATION, MODEL LEARNING AND MAINTENANCE THEORY (DATA-SIMLAMT)

A formal theory for the development of a generic model of an autonomou s sensor is proposed and implemented. An autonomous sensor not only in terprets the acquired data in accordance with an embedded expert syste m knowledge base, but is also capable of using this data to modify and enhance this knowledge base. Hence, the system is capable of learning and thereby improving its performance over time. The main objective o f the model is to combine the capabilities of the physical sensor and an expert operator monitoring the sensor in real time. The system has been successfully tested using various simulated data sets as well as a real thermistor that has been instantiated as an autonomous sensor. Results are provided to show the various tasks associated with the mod el such as sensing, interpretation, maintenance and learning. DATA-SIM LAMT is a novel theory that can make a device capable of human-like re asoning about system behaviors. It finds applications in any field tha t incorporates the human in its control system. Autonomous sensing is but one application of this theory.