FIR system identification under either–or communication with quantised inputs and quantised output observations

• Published in: IET control theory & applications Vol. 13; no. 8; pp. 1204 - 1210
• Main Authors: Guo, Jin, Diao, Jing-Dong
• Format: Journal Article
• Language: English
• Published: The Institution of Engineering and Technology 21.05.2019
• Subjects: binary‐valued output; binary‐valued quantisation; Brief Paper; communication rate; communication resources; either–or communication scheme; empirical measure; event‐triggered output observations; finite impulse response system identification; FIR filters; identification algorithm; identification problem; least squares approximations; multithreshold quantisation; parameter estimation; quantisation (signal); quantised input observations; quantised output observations; quantised‐input FIR system identification; triggered indicator; unknown parameter estimation; vector‐valued observation; weighted least squares; binary-valued quantisation; least squares approximations; multithreshold quantisation; quantised-input FIR system identification; communication resources; unknown parameter estimation; weighted least squares; communication rate; identification algorithm; quantised output observations; either–or communication scheme; vector-valued observation; finite impulse response system identification; FIR filters; quantised input observations; triggered indicator; binary-valued output; event-triggered output observations; empirical measure; quantisation (signal); identification problem; parameter estimation
• ISSN: 1751-8644; 1751-8652; 1751-8652
• DOI: 10.1049/iet-cta.2018.5688

This study investigates the quantised-input FIR (finite impulse response) system identification under quantised and event-triggered output observations. For the binary-valued quantisation, they introduce a so-called either–or communication scheme to save communication resources. Based on the received data and the triggered indicator, an identification algorithm is proposed to estimate the unknown parameter by use of the empirical measure and weighted least squares. The convergence performance is established and the communication rate is derived. For the multi-threshold quantisation, it can be seen as a vector-valued observation in which each component represents a binary-valued output, and the corresponding identification problem is discussed. A numerical example is included to illustrate the main results obtained.