Robust Distributed And Decentralized Control Of Large-Scale Nonlinear Systems With Input Constraints Based On Sos Optimization

This paper studies the robust stabilization problem for large-scale nonlinear systems subject to input constraints and bounded uncertainties. The technical results are developed by using synergies between vector Lyapunov functions and Sum-of-Squares optimization, ultimately enabling the characterization of a subset of the state space from where robust stabilization is guaranteed with the application of subsystem-level, input constrained feedback control laws, designed with Lyapunov techniques. Computationally efficient implementations of such control laws based on lightweight Quadratic Programs are proposed, providing the possibility for either distributed or decentralized system operation, depending on whether information is shared between subsystems or not. Numerical simulations of a complex system illustrate the efficacy of the methods.