Self – Tuning Control of Continuous – Time Systems

One approach to a recursive identification of continuous – time systems was implemented in self – tuning control of a system of interconnected tanks. Since derivatives of input and output variables of continuous – time systems can not be directly measured, differential filters and filtered variables are established to substitute primary variables. The filtered variables are then used in a recursive identification procedure where the classical recursive least squares method is used to identify the system. Results of real – time experiments are compared to results obtained with an analogical discrete controller. The method was also verified by simulation on a multivariable system in order to verify the algorithm on a system with more complex and complicated structure, where one output is influenced by more inputs. Key-Words: self – tuning control, adaptive control, recursive estimation, nonlinear systems, polynomial methods, real – time control