Dynamic Multi-Robot Coalition Formation: Precision Agriculture Case Study

Abstract

The use of multiple autonomous robots to accomplish a complex task is a highly relevant topic for intelligent systems and collective robotics.In this paper, a game-theoretic framework for the self-organization of a group of heterogeneous self-interested robots is described.The proposed approach enables both the tasks allocation and dynamic reward distribution to maximize the expected total gain, which ensures the effectiveness of multirobot coalitions.The solution is based on the theory of fuzzy cooperative games with core.The precision farming scenario is used as an example of a complex task.In this scenario, several robots belonging to two different classes interact with each other to distribute field processing tasks to meet the given marginal cost of each task, corresponding to the coalitions' payoffs.Simulation results show the convergence of the solution when searching for the coalition structures capable of providing a given payoff.That allows to assert the applicability of the theory of fuzzy coalition games for the self-organization in collective robotics.