Robust feedback linearization-based congestion control using a fluid flow model

In this paper, a new robust feedback linearization congestion control strategy for a fluid flow model is introduced. The recourse to a robust control technique permits us to use a non accurate dynamic model in order to design and analyze the controlled system. The fluid flow model (FFM) under its different types of variants is used for network performance evaluation and control as applied to congestion control. Validated by several researchers, the considered first order non linear model is simple in comparison to the detailed Markovian queuing probabilistic models, and it captures the dominant dynamic behavior of a wide range of queuing systems. The sliding mode generalized variable structure (SM-GVS) control recently introduced by M. Fliess which is based on differential algebra concepts allows the switching to take place on the highest derivative of the control input such that the main drawbacks of the discontinuous control that is the chattering is consequently reduced. In this paper, our proposed controller uses the feedback linearization-based SM-GVS approach with some convergence tuning parameters