Using Sub-Micron Spiral Microfluidics and Genetic Programming towards Enhancing Urine-based Glucose Analysis

Improving the accuracy and precision of light spectroscopy-based non-invasive methods for analyzing glucose in urine samples is critical to enabling comfortable, affordable, and robust approaches to diagnosing diabetes. In this paper, we study the feasibility of using spiral microfluidics in studying the separation efficiency in urine simulations. Our results through computational simulations and hardware experiments with sub-micron microfluidics show: (1) the reasonable separation of glucose/urea at 79% glucose separation efficiency, (2) the possibility of relating flow rate and glucose/urea concentration in a closed mathematical form by using Genetic Programming machinery, and (3) the feasibility to render the quasi-linear pattern/relation between glucose concentration and absorbance, which is potential to aid in providing higher/targeted concentrations of glucose for other non-invasive testing approaches, e.g., light spectroscopy.