JORDAN Mario Alberto
congresos y reuniones científicas
Adaptive Robust Control Using Discrete Laguerre Functions
Rio de Janeiro, Brasil
Congreso; 38th IEEE MIDWEST Symposium on Circuits and Systems; 1995
IEEE Circuits and Systems Society
This paper presents a novel approach to adaptive predictive control with robust characteristics based on nonparametric description of linear plants. The approach considers the estimation of a truncated discrete Laguerre series representing the plant to be adaptively controlled. The model uncertainties caused by the series truncation and finite-length data sets are captured by the estimator as tight time-varying upper and lower bounds for the Laguerre coefficients. These map uniquely upper and lower bounds where the actual plant impulse response lies. By combining a FIR model set contained in the family with robust predictive control design techniques in an adaptive control loop bounded stability is ensured. It is shown that that this procedure can alleviate usual strong conditions related to the uncertainty equivalence principle and robust stability can be ensured in a Kharitonov hypercube in the space of controller coefficients. Robust stability conditions and convergence properties of the adaptive control system are performed.