Machine Learning and IoT-Based Ultrasonic Humidification Control System for Longevity of Fruits and Vegetables

Abstract

India wastes approximately Rs 92,651 crores worth of food due to harvest and post-harvest losses. This has an adverse effect on individual farmers who are forced to sell their harvest before it is wasted. The dominant explanation for this situation is the lack of proper storage. Previous research has shown that ultrasonic humidification has the potential to reduce these losses. But it relies on large-scale storage facilities and expensive setup and so the problems of individual farmers remain unresolved. We have boiled down the needs of a proper storage facility into two main factors—temperature and humidity. Contrary to what has been often proposed, we have used a commercial ultrasonic humidifier to create a prototype of a highly economical and portable storage facility. It can automatically optimize the temperature and humidity based on optimal storage conditions for the harvested fruit/vegetable placed inside it which is detected by a machine learning model. The IoT framework also provides visual feedback of the current humidity and temperature data. We have tested the prototype unit against conventional storage methods adopted by farmers. Our results suggest that this prototype unit when developed into a proper storage facility has the potential to reduce the post-harvest losses and the storage issues of individual farmers.