Variable structure-based nonlinear congestion control for networked controlled systems

This paper presents a new robust dynamic congestion control for a network of control systems. Due to an intensive flow of information/data between sensors, decision makers and actuators in the network, synchronization and planning of information/data arrivals, departures, updates and optimal resource allocations take on a significant degree of importance. Sliding mode based variable structure techniques provide robust solutions for complex, nonlinear and strongly coupled control problems and are well-known for their rich analysis and design methodologies. Among the number of possible variable structure algorithms, the proposed controller in this paper is based on the Fliess's approach known as the generalized variable structure (GVS). Introduced recently, this new variable structure approach is designed in the context of differential algebra. Containing principally the derivative of the input, the new observable canonical form is used for design of the controller. Thereby facilitating the main benefit of this algorithm as far as reducing the correspondingly known chattering phenomena that is the main drawback in standard variable structure control (VSC) techniques.