The digital cold chain: Sensor-driven product quality with AI
Rania Elashmawy, Moshe Doron, Ria Kanjilal, Jeffrey K. Brecht, Ismail Uysal
Abstract
The concept of using data-driven computational models to represent a physical object or process in a digital environment, such as the cloud, has become increasingly practical with the ubiquitous application of sensors and other IoT devices. For instance, the shelf life of a strawberry can be predicted throughout the entire distribution process based on the initial quality, temperature, and shipment duration. The marketability and quality of strawberries, including color, sugar content, and firmness, deteriorate during the postharvest process. Accurate prediction of these metrics can facilitate maintaining quality standards during distribution and enable smart distribution. In this article, we introduce a sensor-driven AI/ML-enabled algorithm to provide insight into the digital cold chain of strawberries and demonstrate that using simple sensor data, can create reliable representations of the marketability, color, sugar content, and firmness of a digital strawberry. Marketability and sweetness ratio are predicted with error percentages of 4.21 (%) and 15.56 (%) within the expected range of values, respectively, with support vector regression. Decision tree regressions achieved prediction error percentages of 1.95 (%) and 4.12 (%) within the expected range of values for color and firmness, respectively. Ultimately, a first-expired-first-out distribution chain can replace the industry standard first-in-first-out to prevent loss with the use of the proposed methods for an accurate and validated digital cold chain.