Investigation of Harmful Substances to Health Using Electronic Nose and Weight Elimination Algorithm With Correlative Gaussian Function

Humans can be exposed to various substances that could either contribute, cause, or catalyze illness. Electronic Nose can be used to detect presence of such elements in excess in human body, through smelling of human breath, skin, or body fluid samples. The objective of this project is to implement a simple, low cost, and portable E-Nose that can be used to detect substances, such as, acetone, ammonia, alcohol, which contribute to Diabetes and Kidney Failure, among other elements, which could cause illness, or support further effect of the illness on human daily life. The main feature of the presented nose in this work, which uses Arduino hardware, is the intelligent Neural Networks software with Correlative Gaussian interpolation function. This function is used together with the Weight Elimination Algorithm (WEA) to enable smart classification of detected substances. The WEA works similar to genetic algorithm in the way it eliminates weak weights and links. Together with Correlative Gaussian, a Correlative Weight Elimination Algorithm is produces (CWEA). Such intelligent discrimination technique allowed not only to detect and classify chemicals in a single substance, but also, to detect and classify the same element and its overall effect in multiple substances. The obtained results are promising, with better results, and possibility of covering more substances, is applicable using higher level and integrated sensors.