The control of pure feedback system, which is widely used but has non-affine property, has always been an important and challenging problem. In order to achieve precise tracking control of pure feedback system through improving the disturbance rejection ability of existing reinforcement learning algorithm, a reinforcement learning (RL) control strategy based on extended state observer (ESO) is proposed in this paper. In the proposed method, the extended state observer can reject the total disturbances and transform the pure feedback system which is in an input-output predictor from to overcome the non-causal problem into a cascade integral form. This allows the continuous reinforcement learning strategy of the actor-critic (AC) structure not to depend on the detailed model information, which makes it practically data-driven. It is worth noting that, in order to further improve the ability to track the changing reference trajectory, a novel curvature acceleration factor is proposed, which can adjust the learning speed of the reinforcement learning controller according to the curvature of the reference trajectory. The validity of the proposed algorithm is verified by the simulation results.
Skip Nav Destination
ASME 2017 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference
August 6–9, 2017
Cleveland, Ohio, USA
Conference Sponsors:
- Design Engineering Division
- Computers and Information in Engineering Division
ISBN:
978-0-7918-5823-3
PROCEEDINGS PAPER
On the ESO Based Reinforcement Learning for Pure Feedback Systems
Dazi Li,
Dazi Li
Beijing University of Chemical Technology, Beijing, China
Search for other works by this author on:
Wei Wang,
Wei Wang
Beijing University of Chemical Technology, Beijing, China
Search for other works by this author on:
Zhiqiang Gao
Zhiqiang Gao
Cleveland State University, Cleveland, OH
Search for other works by this author on:
Dazi Li
Beijing University of Chemical Technology, Beijing, China
Wei Wang
Beijing University of Chemical Technology, Beijing, China
Zhiqiang Gao
Cleveland State University, Cleveland, OH
Paper No:
DETC2017-67659, V009T07A019; 6 pages
Published Online:
November 3, 2017
Citation
Li, D, Wang, W, & Gao, Z. "On the ESO Based Reinforcement Learning for Pure Feedback Systems." Proceedings of the ASME 2017 International Design Engineering Technical Conferences and Computers and Information in Engineering Conference. Volume 9: 13th ASME/IEEE International Conference on Mechatronic and Embedded Systems and Applications. Cleveland, Ohio, USA. August 6–9, 2017. V009T07A019. ASME. https://doi.org/10.1115/DETC2017-67659
Download citation file:
53
Views
Related Proceedings Papers
On the ADRC of Non-Differentially Flat, Underactuated, Nonlinear Systems: An Experimental Case Study
IDETC-CIE2017
Tracking Control of an Underwater Vehicle Using Adaptive Back-Stepping With Tuning Function
IDETC-CIE2005
Related Articles
On the Linear Control of Underactuated Nonlinear Systems Via Tangent Flatness and Active Disturbance Rejection Control: The Case of the Ball and Beam System
J. Dyn. Sys., Meas., Control (October,2016)
Tracking Control of Limit Cycle Oscillations in an Aero-Elastic System
J. Dyn. Sys., Meas., Control (November,2014)
Real Time Estimation of Elastic Deformation for End-Point Tracking Control of Flexible Two-Link Manipulators
J. Dyn. Sys., Meas., Control (September,1993)
Related Chapters
A Novel Frequency Domain Equalization Method for PN-Based Single-Carrier System
International Conference on Mechanical and Electrical Technology, 3rd, (ICMET-China 2011), Volumes 1–3
Distributed Traffic State Estimation and Classification Using Consensus-Based Expectation Maximization Algorithm in Spatially Deployed Traffic Detectors
International Conference on Mechanical Engineering and Technology (ICMET-London 2011)
An Adaptive Fuzzy Control for a Multi-Degree-of-Freedom System
Intelligent Engineering Systems Through Artificial Neural Networks, Volume 17