Developing a smart cyber-physical system based on digital twins of plants

The paper proposes a multi-agent approach to development of “digital twins” of plants, which reflects phases of plant development and allows for more accurate forecasting of harvest and planning of agrotechnical measures. It also shows the possibility to formalize domain knowledge on new agro technologies for plant growing and automate decision-making processes when introducing precision farming technologies. A formalized model of the digital twin of plants is proposed in the form of a multi-agent state machine, acting on the basis of a knowledge graph of transitions between phases of plant development. Each stage (state) has its own software agent, which helps respond to unforeseen events and recalculate predicted harvest and risks of problem situations. Recommendations developed by the digital twin are to be further sent to smart cyber-physical precision farming management system for adaptive rescheduling of resources and differentiated use of fertilizers and crop protection tools. The developed approach becomes valuable in growing uncertainty because of global climate changes, destroying agronomist domain knowledge collected over centuries.