Using a Kalman filter and a Pade approximation to estimate random time delays in a networked feedback control system

This study presents an effective algorithm to handle time delays in a feedback control loop in which both the measurement signals and the control signals are sent through a network where random time delays and packet loss occur. A method is developed to estimate the random time delay when no time stamps are present. Specifically, a Padé approximation is used to model time delay in the frequency domain, which can then be estimated by a Kalman filter. The method is applied to an inverted pendulum, where its effects are illustrated.