Improved intelligent identification of uncertainty bounds: design, model validation and stability analysis

Abstract

Identification of uncertainty bounds in robust control design is known to be a critical issue that attracts the attention of research in robust control field recently. Nevertheless, the practical implementation involves a trial and error procedure, which depends on the designer prior knowledge and the available information about the system under study. Artificial intelligent techniques provide a suitable solution to such a problem. In this paper a new intelligent identification method of uncertainty bound using an adaptive neuro-fuzzy inference system (ANFIS) in an enhanced feedback scheme is proposed. The proposed ANFIS structure enables accurate determination of the uncertainty bounds and guarantees robust stability and performance. In our proposed technique, the validation of the intelligent identified uncertainty weighting function is based on the measurement of both the v-gap metric and the stability margin that result from the corresponding robust controller design. Additionally, these two indices are used to improve the accuracy of the intelligent estimation of uncertainty bound in conjunction with the robust control design requirements. The enhanced intelligent identification of uncertainty bound is demonstrated on a servo positioning system. Simulation and experimental results prove the validity of the applied approach; more reliable and highly efficient estimation of the uncertainty weighting function for robust controller design.