BYU

Abstract by Charles Lewis

Personal Infomation


Presenter's Name

Charles Lewis

Degree Level

Undergraduate

Abstract Infomation


Department

Physics and Astronomy

Faculty Advisor

John Colton

Title

Neural Network Approximations for CdTe Temperature Sensors

Abstract

Various biological processes require accurate temperature sensing through microfluidic devices, so improved temperature sensors methods are needed. Previously, machine learning techniques have been used for predicting temperatures through thermal images. To develop a new, accurate temperature sensor model, photoluminescence (PL) spectra and time-resolved photoluminescence (TRPL) spectra of CdTe quantum dots were measured as functions of temperature for use in training an artificial neural network (ANN). A low temperature regime from 10-320 K and a high temperature regime from 325-346 K were measured with additional data provided through interpolation. The optimized neural network is able to determine temperatures with a mean average error of 5.2 K and 0.2 K for the low and high temperature regimes respectively.