This work demonstrates improved system identification using a state vector and ARMA modeling in linear time varying systems, implying enhanced control capabilities.
Key Points
To develop a system identification algorithm for linear time-varying systems using an information-state approach.
Proposes a system realization approach based on an information-state vector.
Uses input-output data to fit an autoregressive moving average model (ARMA).
Details the theoretical foundation for ARMA-based system representation using linear observability.
Assesses performance on various systems compared to state-of-the-art techniques.
Establishes a direct realization of state-space models using estimated time-varying ARMA parameters.
Validates the effectiveness of the approach in optimal output feedback control scenarios.
Shows no need for separating free and forced response for identification.