Adaptive identification for self-tuning controllers design

The analysis and implementation of an adaptive identification method for the tuning of self-tuning controllers is presented. Real-time identification of the plant to be controlled allows online availability of its parameters to tune a digital controller via the method of dynamic assignment of poles and zeros in a "Self Tuning Regulator (STR)" control configuration. The operation of this proposal is based on the execution of two consecutive processes: on one hand, identification of the plant through estimation; and on the other hand the subsequent calculation of controller's set of poles and zeros. These tasks are executed between successive instants of sampling, achieving in this way real time self-tuning of the controller. Starting from the algorithmic formulation of the proposed identification method besides its computational simulation, functional verification was carried out through the design and development of a DSP platform, which fulfills didactic and research needs in academic subjects like Control Systems and Digital Signal Processing. In this way, a system for self-tuning of a digital controller is available, using the RLS algorithm for plant identification. The operation of the system was verified using known second and third order plants, obtaining results that confirm correct operation of the system as described in the text of this work.