Challenge of intelligent sensing for monitoring and control

Intelligent sensors offer new capabilities in process measurement, including networked communications, self-diagnostics, sampling on-demand, and dynamically defined measurement uncertainty intervals. Traditional control and system theory models do not adequately consider the measurement uncertainty information and the nonperiodic sampling possible from intelligent sensors. This paper introduces a new modeling framework for systems exhibiting these irregular observations. The framework uses linear constraints and a variation of descriptor system models to represent relational dynamics and set-valued observation information. The paper illustrates the use of the model framework for set-valued state estimation.