Neutral prediction in control of time-delay systems

Controller gain is often limited by time delay in the control loop and therefore actuation is constrained and the steady-state error is considerable due to frictions existing. Consideration is given to dynamic modelling and positioning of a pneumatic servo-actuator with time delay and friction. A Smith predictor-type control scheme is considered to improve the control behaviour which is incorporated with a Gain Scheduling method aiming to reduce steady-state error. A type of artificial neural network, Group Method and Data Handling (GMDH), is used for model prediction. Results show that the time delay effect is removed satisfactorily, systems stability significantly improved, response speed raised and steady-state error reduced