Decentralized Supervision for the Cooperation of Heterogeneous Field Robots in an Orchard

Heterogeneous field robot systems are large-scale and dynamic. The use of such robot systems for system modeling and controller design is challenging, but these systems are being actively studied in various research fields. In this study, we model a heterogeneous field robot system based on hybrid automata for the cooperation of heterogeneous field robots. We also propose a decentralized supervisory control architecture for the modular control of the field robots. The performance of heterogenous robots in our proposed system is more efficient and systematic than that of classical control methods for managing heterogeneous systems. Furthermore, we identify that it is possible to analyze not only the continuous-time dynamics of the system but also the discrete-event dynamics. To evaluate the proposed system, experiments were conducted using a physics-based simulator. The experimental results indicated that our decentralized supervisory control system realized the cooperation of heterogeneous field robots based on the designed constraints.