Prescribed finite time consensus of networked multi-agent systems

Abstract

The paper presents a finite-time distributed control method for consensus of networked multiple systems, which is different from the traditional methods based on signum function or fractional power state feedback (where the finite convergence time is contingent on initial conditions and the control action is discontinuous or non-smooth). More specifically, the proposed method is built upon the regular state feedback, incorporated with a finite-time scaling function, leading to distributed smooth control action. Furthermore, with this method, the consensus is achieved within prescribed-time under bidirectional interaction. Namely, all the agents reach the average consensus in designer-assigned finite time under undirected connected topology. Numerical simulations demonstrate and validate the superiority of the proposed control.