Data-driven Search for the Graph Structure Improving Control Performance in Cooperative Control of Human-vehicle Network

Abstract

Recently, robot swarm systems have been developed and cooperative control of human-robotic network based on passivity of a human operator has been proposed. However, it is difficult to guarantee control performance of human-robotic network because it depends on the property of human. In this paper, we propose a data-driven method to search for the graph structure of the vehicle swarm improving control performance in cooperative control of human-vehicle network, assuming that humans are unknown linear systems. We use the idea of FRIT and define a cost function which means the error between the output of the ideal transfer function and the vehicle swarm. Then, we propose a method to calculate the cost function with an one-shot input output data. In this way, we can search for the graph structure of the vehicle swarm that is easy for humans to control. Finally, we demonstrate the effectiveness of the proposed method through numerical simulations.