Moving Toward Self-Learning Closed Plant Production Systems

Abstract

A general scheme of next-generation closed plant production systems (CPPS) with self-learning ability (s-CPPS), consisting of virtual and real CPPS as the two main components, is presented. In virtual CPPS, various models to show the environmental effects on plant growth and mass and energy balance in the CPPS are embeded and then used for predictive simulations of real CPPS. The values of coefficients (parameters) in the models are adjusted by a parameterization process to fit better with those in real CPPS. Once an s-CPPS is developed, it can be utilized, after some revisions, in open and semi-closed greenhouse crop production systems. Recent global technologies are advantageous in s-CPPS. For example, information and communication technology (ICT); artificial intelligence (AI) with big data mining, deep learning, pattern recognition, etc.; the Internet of Things (IoT); and bioinformatics can be combined to make s-CPPS smarter. s-CPPS are integrated plant production systems for next-generation urban agriculture.