Abstract
The paper is concerned with a hybrid approach to the design of parameter-adaptive control systems combining continuous-time process models with discrete-time control. Two different sampling rates are introduced for process identification and control in order to avoid numerical problems and to increase control performance respectively.Particular emphasis is laid on the identification of continuous-time process models. A brief analysis of the state variable filter approach shows that the resulting parameter estimates are incorrect in case digital filters are used. Therefore an improved state variable filter is introduced which reduces the estimation errors substantially by means of interpolation techniques. The resulting filter algorithm is quite easy to implement and suitable for on-line process identification even in case of fast sampling. Based on the continuous-time process model a digital PID-controller is designed by minimizing a quadratic cost functional and discretizing the resulting differential equation of the continuous-time controller.The performance of the hybrid controller is demonstrated by some experimental results.
Published Version
Talk to us
Join us for a 30 min session where you can share your feedback and ask us any queries you have
Schedule a call