Recursive least determinant self-tuning regulator
An algorithm for the self-tuning regulator of a Box-Jenkins model control system is suggested. The algorithm calculates the control action by minimising the determinant of a positive definite matrix formed by the values of the input and output variables. The algorithm will gradually reduce a sequence of these matrices to a singular matrix. At this condition, the parameters of the regulator can be obtained from an eigenvector corresponding to the zero eigenvalue of this matrix. The algorithm does not need to know the delay of the process dynamics which is its strength over the algorithm of the recursive least squares self-tuning regulator algorithm.