Control architecture for heterogeneous multiple robots with human-in-the-loop

Multi-robot system often require human assistance in practical situation. The level of controlling multiple robots in the real world lies somewhere between fully autonomous and fully manual. The optimal level of autonomy for the multi-robot system controller is discussed in this paper. Since human perception is limited for controlling large number of robots, interface design should consider separating autonomous and manual part. Therefore, this paper proposes a control architecture for multi-robot system enabling human-in-the-loop. The architecture is composed of three layers: supervisory, coalition, and individual control layer. The supervisory control layer utilizes human engagement, the coalition control layer decides the cooperative task of robots, and the individual control layer controls the motion of robot autonomously.