Intelligent sensor analysis and actuator control

Abstract

This paper describes a tool called Isaac (intelligent sensor analysis and actuator controller) that autonomously explores the behavior of a dynamical system and uses the resulting knowledge to help build and test mathematical models of that system. Isaac is a unified knowledge representation and reasoning framework for input/output modeling that can be incorporated into any automated tool that reasons about dynamical models. It is based on two modeling paradigms, intelligent sensor data analysis and qualitative bifurcation analysis, which capture essential parts of an engineer’s reasoning about modeling problems. We demonstrate Isaac’s power and adaptability by incorporating it into the Pret automated system identification tool and showing how input/output modeling expands Pret’s repertoire.