Iterative learning parameter estimation and design of SMB processes

Learning speed enhancement is one of the most important issues in learning control. If we can improve both learning speed and tracking performance, it will be helpful to the applicability of learning control. Considering these facts, in this paper, we propose a learning speed enhancement scheme for iterative learning control with advanced output data (ADILC) based on parameter estimation. We consider linear discrete-time non-minimum phase (NMP) systems, whose model is unknown, except for the relative degree and the number of NMP zeros. In each iteration, estimates of the impulse response are obtained from input-output relationship. Then, learning gain matrix is calculated from the estimates, and by using new learning gain matrix, learning speed can be enhanced. Simulation results show that the learning speed has been enhanced by applying the proposed method.

Download Full-text

Machine Learning Parameter Estimation in a Smart-City Paradigm for the Medical Field

Smart Cities Performability, Cognition, & Security - EAI/Springer Innovations in Communication and Computing ◽

10.1007/978-3-030-14718-1_7 ◽

2019 ◽

pp. 139-151 ◽

Cited By ~ 1

Author(s):

M. Bhuvaneswari ◽

G. Naveen Balaji ◽

F. Al-Turjman

Keyword(s):

Machine Learning ◽

Parameter Estimation ◽

Smart City ◽

Medical Field ◽

Learning Parameter

Download Full-text

Revisiting Boltzmann learning: parameter estimation in Markov random fields

1996 IEEE International Conference on Acoustics, Speech, and Signal Processing Conference Proceedings ◽

10.1109/icassp.1996.550606 ◽

2002 ◽

Cited By ~ 1

Author(s):

L.K. Hansen ◽

L.N. Andersen ◽

U. Kjems ◽

J. Larsen

Keyword(s):

Parameter Estimation ◽

Random Fields ◽

Markov Random Fields ◽

Markov Random ◽

Learning Parameter

Download Full-text

Nonlinear Piezo-Actuator Control by Learning Self-Tuning Regulator

Journal of Dynamic Systems Measurement and Control ◽

10.1115/1.2899203 ◽

1993 ◽

Vol 115 (4) ◽

pp. 720-723 ◽

Cited By ~ 32

Author(s):

C. James Li ◽

Homayoon S. M. Beigi ◽

Shengyi Li ◽

Jiancheng Liang

Keyword(s):

Parameter Estimation ◽

Tracking Error ◽

Diamond Turning ◽

Piezo Actuator ◽

Positioning System ◽

Tool Positioning ◽

Repetitive Tasks ◽

Self Tuning ◽

Learning Parameter ◽

Actuator Control

This paper presents a learning self-tuning (LSTR) regulator which improves the tracking performance of itself while performing repetitive tasks. The controller is a self-tuning regulator based on learning parameter estimation. Experimentally, the controller was used to control the movement of a nonlinear piezoelectric actuator which is a part of the tool positioning system for a diamond turning lathe. Experimental results show that the controller is able to reduce the tracking error through the repetition of the task.

Download Full-text