An improved algebraic geometric solution to the identification of switched ARX models with noise

In this paper, we present an improved algebraic geometry solution for the identification of switched ARX models in the presence of measurement noise. The procedure utilizes the highest order of sub-models, which is estimated by using statistical analysis of effective singular values in matrix rank determination. After embedding sub-models into a large continuous-time model for omitting the necessity of switching sequence, an analytical solution for the two-mode system is obtained using matrix differential calculus. The improvements made to the previous method are verified by simulations on two linear systems. Also the effectiveness of the proposed method is shown by using a two mode experimental pilot plant.