AMBAR is a mobile and web application that helps apple cold storage staff see the risk of condensation (water forming on the apple skin) before it causes decay. Most cold rooms only measure air temperature and humidity, so the real temperature of the fruit is not known. AMBAR uses these air measurements and machine learning models, trained on a real apple cold storage dataset, to estimate the condensation risk for every sensor, bin stack and tier in the room. Storage staff follow the live risk levels on their phones, receive push notifications when the risk becomes high and look at the history and charts to understand what happened. Managers set alert thresholds and export reports, and the models can be evaluated and updated from the web application.
