SmartSniffer: Predicting Food Spoilage Time with an Electronic Nose-Based Gas Monitoring Apparatus Utilizing a Two-Stage Pipeline Model

Food waste is a growing concern in the sustainable development goal (SDG) standards accepted in the world. Hence reducing food wastage and timely consumption of the healthy food is socially and economically appreciated. Moreover, the consumption of non-edible food not only poses a prolonged health risk but, more significantly, the d i agnosis o f illness is not immediately apparent. Hence detection and preferably prediction of food spoilage needs undivided attention from the food man-ufacturers. However various foods have different shelf life and different appearances on spoilage makes it extremely difficult to identify on the edibility of the food item. Since it's not possible to taste and measure the freshness of food items especially in large setups of food distribution systems, it is beneficial to have an electronic setup to identify the same. Many research works have used gas sensors along with Neural Network to detect the food spoilage. However estimating the time of the spoilage will not only enable the food consumption at appropriate time, but also help in regulating the food wastes. In the proposed approach, the time of spoilage is predicted using time-series analysis model and Artificial Neural Network. This novel approach has yielded an accurate estimation of spoilage time making it a potential solution for management of food items and advice for consumption during their freshness. This work is a step towards realizing a better food management and supply system with an aim to reduce the wastage and at the same time consume healthy food. The two-stage pipeline model with the dataset is made publicly available for easy adoption and further usage by the researchers and developers community.