CONSTRAINED DISCRETE-TIME STATE-DEPENDENT RICCATI EQUATION CONTROL FOR DECENTRALIZED MULTI-AGENT SYSTEMS

The objective of this paper is to introduce the constrained discrete-time state dependent Riccati equation (CD-SDRE) technique for decentralized multi-agent systems. First, a D-SDRE for tracking control is derived analytically. Then, the latter is utilized in formulating the distributed D-SDRE control law. The constraint problem of the D-SDRE/distribued D-SDRE is then addressed by recasting it into a model predictive control (MPC) problem. Solution of the latter is then carried out using particle swarm optimization (PSO), which facilitates estimation of the states and control inputs for the D-SDRE control law calculation. The use of decentralized CD-SDRE scheme in a networked system in the presence of constraints is then systematically demonstrated by applying this scheme to reconfiguration problems of spacecraft formation flying under constrained actuation. Simulation results demonstrate the efficacy and reliability of the proposed CD-SDRE in the decentralized control of the spacecraft formation flying.