Application of Maximum Hands-off Distributed Control to a Quadrotor Group

In a multi-agent system, all agents achieve one task as a group with the own decision of each agent. In this research, we consider quadrotors as agents and investigate the consensus control method to converge the states of all quadrotors to the same state (reach a consensus). In the consensus control, each quadrotor should calculate the control input by itself using the limited information about other quadrotors. Here, we assume a distributed control system and the agents can only observe the state of specific agents. To solve this consensus problem, we introduce the maximum hands-off distributed control based on the sparse optimal control. In the maximum hands-off distributed control, we calculate not only the control input to reach a consensus but also the sparse control input in the time horizon direction and maximize the duration in which the control inputs are exactly zero. Therefore, minimizing the necessary energy for control is expected. In this paper, we apply the maximum hands-off distributed control to a case in which four quadrotors move as a group and confirm that the quadrotors can reach consensus in the simulations and actual experiments using only limited information.