Bayesian Design Optimization of a Non-specific Sensor System for Calibration of Analyte Responses
David Han, University of Texas at San Antonio, Research Interests: reliability engineering, survival analysis, statistical applications
In the current and future generation of products, the nature of field reliability data is changing rapidly and dramatically. With modern sensor technology, innovative data analytics is emerging in reliability and quality technology. In order to reduce unexpected failures of a system in service, it is crucial to assess its condition/health accurately in real time. Using an array of sensors with well calibrated but different tuning curves, it is possible to appreciate a wide range of stimuli. The objective of this research is to adopt Bayesian framework to develop a statistically sound estimation method given sensor responses by elucidating the uncertain nature of environment-dependent stimuli through a choice of prior. Using decision-theoretic approach, the design optimization of a sensory system is also explored via maximization of the expected utility. Furthermore, to characterize the fundamental analytical capability of a measurement system, general-purpose selectivity is defined in an information-theoretic manner.