A data-driven controllability measure for linear discrete-time systems

An data-based approach is developed to analyze the controllability of linear discrete-time systems in this paper, in which the parameter matrices are unknown. The proposed method only applies the measured state and output data to verify the system property, instead of the process data of the system mathematical model. The direct analysis method can be used and the unknown parameter matrices are not necessary to gain. The data-based methods can avoid identification errors and have lower computational complexity than the traditional model-based analysis methods. Controllability measurement set forth above not only shows if the system is controllable or not, but also reveals the level of controllability of the system.