An AIOT-based Evaluator Design for Sesame Oil

Sesame oil is a commonly used as a seasoning oil. The edible oil quality is usually determined by chemical indexes, including acid value, peroxide value and p-anisidine value. It is difficult for a user to test whether a sesame oil is bad. Hence, this study designs an AIOT-based Sesame Oil Evaluator, which is called ASOE, composed of ten gas sensors and a machine learning inference engine. Ten gas sensors sense ethanol, methane, hydrogen sulphide, and ammonia. Sensed data is analyzed by the Linear Discriminant Analysis (LDA). Evaluation results indicate that the correction rate of identifying one to nine week-old sesame oil by LDA is 86%. ASOE can help a user easily and quickly evaluate the quality of a sesame oil to substitute chemical indexes. It can be a reference model for developing a food flavoring oil detector in the future.