Integrating Curing Experience and Fluctuating Leaf Quality in an Intelligent End-Cloud Collaboration

Abstract

Automated tobacco curing systems have been widely deployed to meet the demand of large-scale production of high-quality tobacco. Existing tobacco curing systems control the process based on a fixed standard curing curve to increase the production of high-quality tobacco and reduce the cost. However, it is difficult to adapt the standard curing curve to the various qualities of tobacco leaves and different experience of operators. In this paper, we propose an end-cloud collaborative intelligent curing technique which can continuously extract and integrate curing experience by using clustering and regression analysis method. At the terminal, we use a microcontroller to receive sensor data collected and exchange the information with servers through a 4G communication module. On the server side, we use a clustering-based method to separate the data of high-quality tobacco, and use regression analysis and kalman filtering for data fusion. Compared with the existing automatic curing method, our proposed method can continuously extract curing experience and optimize curing curves.