A distributed controller approach for delay-independent stability of networked control systems

This article introduces a novel distributed controller approach for networked control systems (NCS) to achieve finite gain L2 stability independent of constant time delay. The distributed controller is alternatively interpreted as a linear transformation of the inputoutput space of the controller and the plant, and in fact represents a generalization of the well-known scattering transformation. The main results of this article are a) a sufficient stability condition for general multi-input-multi-output (MIMO) input-feedforward-outputfeedback-passive (IF-OFP) nonlinear systems and b) a necessary and sufficient stability condition for linear time-invariant (LTI) single-input-single-output (SISO) systems. The performance advantages in terms of sensitivity to time delay and steady state error are discussed in comparison with alternative delay-independent small gain type approaches. Simulations verify the superiority of the proposed approach to an LQR controller for zero time delay without the proposed transformation and to a delay-independent small gain based controller.