State-Space System Identification with Identified Hankel Matrix

In state-space system identification theory, the Hankel matrix often appears prior to model realization. Traditionally, one identifies from input-output data the Markov parameters from which the Hankel matrix is built. This paper examines the strategy where the Hankel matrix itself is identified from input-output data. Various options are examined along this direction where the identification of the Hankel matrix can be carried out directly or indirectly. Advantages and drawbacks associated with each option are examined and discussed. Extensive evaluation both with simulated and experimental data indicates that the new approach is effective in detecting the “true” or effective order of the system, hence it is capable of producing relatively low-dimensional state-space model. The interactive matrix formulation plays a key role in the development of the proposed identification technique.