Design of Robust Stochastic Controller for Stabilizing Queues of TCP/AQM System

In this article, a robust stochastic controller is designed analytically using the linearized fluid flow model of TCP/AQM behavior. The stability criteria found is delay dependent. The control law is obtained as a solution of linear matrix inequalities (LMIs) and not only asymptotically stabilizes the queue size to the desired operating point, but also guarantees the robust stability of network against the linearization error in estimating round-trip time delay. The performance of the proposed controller is verified and compared with the PI and ARED AQM schemes using the NS-2 simulator. The simulation results show that the proposed scheme has better performance and robustness than PI and ARED controllers with faster response, smaller overshoot and less bias.