Distributed observer-based LQR design for multi-agent systems

In this paper, the problem of designing distributed LQR for multi-agent systems is considered when the states are not available for the control purpose. Two types of distributed observer-based LQR controller are developed with local observers and distributed observers. The design procedure provides a systematic way to construct control and observer gains that guarantee stabilization by solving two low-dimensional algebraic Riccati equations (ARE). The control and observer gains are designed in separation mode and they are decoupled from the communication graph structure. The effectiveness of this method is demonstrated by an example.