Multiple model approach by orthonormal bases for controller design

Recently, model-free or data-driven control has been gaining great interest in overcoming the limitations of conventional model based control methodologies. However, the existing data-driven control is far from practical because of its slow convergence, severe computational burden and lack of analysis and synthesis tools. This research was motivated in response to these deficiencies of data-driven control methodologies by multiple model approach. In this paper, a multiple modeling method for subsequent robust controller design is developed. This method is developed to overcome the difficulties of utilizing the existing multiple model approaches for control purpose by adopting a different locality concept based on dominant poles. The dominant poles are represented by Laguerre basis functions. An example is included to demonstrate the use and characteristics of the proposed as well as existing identification methods