Title: Distributed Optimal Control with Recovered Robustness for Uncertain Network Systems: A Complementary Design Approach

Abstract: This paper considers the distributed robust suboptimal consensus control problem of linear multi-agent systems, with both H2 and H_infty performance requirements. A novel two-step complementary design approach is proposed. In the first step, a distributed control law is designed for the nominal multi-agent system to achieve consensus with a prescribed H2 performance. In the second step, an extra control input, depending on some carefully chosen residual signals indicating the modeling mismatch, is designed to complement the H2 performance by providing robustness guarantee in terms of H_infty requirement with respect to disturbances or uncertainties. The proposed complementary design approach provides an additional degree of freedom for design, having two separate controls to deal with the H2 performance and the robustness of consensus, respectively. Thereby, it does not need to make much trade-off, and can be expected to be much less conservative than the trade-off design such as the mixed H2/H_infty control method. Besides, this complementary approach will recover the achievable H2 performance when external disturbances or uncertainties do not exist. The effectiveness of the theoretical results and the advantages of the complementary approach are validated via numerical simulations.