Neural Identification for Control

Neural Identification for Control. Saha, P. & Mukhopadhyay, S. arXiv:2009.11782 [cs, eess], September, 2020. arXiv: 2009.11782

Paper abstract bibtex

We present a new method for learning control law that stabilizes an unknown nonlinear dynamical system at an equilibrium point. We formulate a system identification task in a self-supervised learning setting that jointly learns a controller and corresponding stable closed-loop dynamics hypothesis. The open-loop input-output behavior of the underlying dynamical system is used as the supervising signal to train the neural network-based system model and controller. The method relies on the Lyapunov stability theory to generate a stable closed-loop dynamics hypothesis and corresponding control law. We demonstrate our method on various nonlinear control problems such as n-Link pendulum balancing, pendulum on cart balancing, and wheeled vehicle path following.

@article{saha_neural_2020,
	title = {Neural {Identification} for {Control}},
	url = {http://arxiv.org/abs/2009.11782},
	abstract = {We present a new method for learning control law that stabilizes an unknown nonlinear dynamical system at an equilibrium point. We formulate a system identification task in a self-supervised learning setting that jointly learns a controller and corresponding stable closed-loop dynamics hypothesis. The open-loop input-output behavior of the underlying dynamical system is used as the supervising signal to train the neural network-based system model and controller. The method relies on the Lyapunov stability theory to generate a stable closed-loop dynamics hypothesis and corresponding control law. We demonstrate our method on various nonlinear control problems such as n-Link pendulum balancing, pendulum on cart balancing, and wheeled vehicle path following.},
	urldate = {2020-09-28},
	journal = {arXiv:2009.11782 [cs, eess]},
	author = {Saha, Priyabrata and Mukhopadhyay, Saibal},
	month = sep,
	year = {2020},
	note = {arXiv: 2009.11782},
	keywords = {machine learning, mentions sympy, robotics},
}

Downloads: 0

{"_id":"2sW5RyESuMcQ7mHNq","bibbaseid":"saha-mukhopadhyay-neuralidentificationforcontrol-2020","authorIDs":[],"author_short":["Saha, P.","Mukhopadhyay, S."],"bibdata":{"bibtype":"article","type":"article","title":"Neural Identification for Control","url":"http://arxiv.org/abs/2009.11782","abstract":"We present a new method for learning control law that stabilizes an unknown nonlinear dynamical system at an equilibrium point. We formulate a system identification task in a self-supervised learning setting that jointly learns a controller and corresponding stable closed-loop dynamics hypothesis. The open-loop input-output behavior of the underlying dynamical system is used as the supervising signal to train the neural network-based system model and controller. The method relies on the Lyapunov stability theory to generate a stable closed-loop dynamics hypothesis and corresponding control law. We demonstrate our method on various nonlinear control problems such as n-Link pendulum balancing, pendulum on cart balancing, and wheeled vehicle path following.","urldate":"2020-09-28","journal":"arXiv:2009.11782 [cs, eess]","author":[{"propositions":[],"lastnames":["Saha"],"firstnames":["Priyabrata"],"suffixes":[]},{"propositions":[],"lastnames":["Mukhopadhyay"],"firstnames":["Saibal"],"suffixes":[]}],"month":"September","year":"2020","note":"arXiv: 2009.11782","keywords":"machine learning, mentions sympy, robotics","bibtex":"@article{saha_neural_2020,\n\ttitle = {Neural {Identification} for {Control}},\n\turl = {http://arxiv.org/abs/2009.11782},\n\tabstract = {We present a new method for learning control law that stabilizes an unknown nonlinear dynamical system at an equilibrium point. We formulate a system identification task in a self-supervised learning setting that jointly learns a controller and corresponding stable closed-loop dynamics hypothesis. The open-loop input-output behavior of the underlying dynamical system is used as the supervising signal to train the neural network-based system model and controller. The method relies on the Lyapunov stability theory to generate a stable closed-loop dynamics hypothesis and corresponding control law. We demonstrate our method on various nonlinear control problems such as n-Link pendulum balancing, pendulum on cart balancing, and wheeled vehicle path following.},\n\turldate = {2020-09-28},\n\tjournal = {arXiv:2009.11782 [cs, eess]},\n\tauthor = {Saha, Priyabrata and Mukhopadhyay, Saibal},\n\tmonth = sep,\n\tyear = {2020},\n\tnote = {arXiv: 2009.11782},\n\tkeywords = {machine learning, mentions sympy, robotics},\n}\n\n\n\n\n\n\n\n\n\n\n\n","author_short":["Saha, P.","Mukhopadhyay, S."],"key":"saha_neural_2020","id":"saha_neural_2020","bibbaseid":"saha-mukhopadhyay-neuralidentificationforcontrol-2020","role":"author","urls":{"Paper":"http://arxiv.org/abs/2009.11782"},"keyword":["machine learning","mentions sympy","robotics"],"metadata":{"authorlinks":{}},"downloads":0},"bibtype":"article","biburl":"https://bibbase.org/zotero-group/nicoguaro/525293","creationDate":"2020-10-02T10:13:31.586Z","downloads":0,"keywords":["machine learning","mentions sympy","robotics"],"search_terms":["neural","identification","control","saha","mukhopadhyay"],"title":"Neural Identification for Control","year":2020,"dataSources":["YtBDXPDiQEyhyEDZC","fhHfrQgj3AaGp7e9E","qzbMjEJf5d9Lk78vE","45tA9RFoXA9XeH4MM","MeSgs2KDKZo3bEbxH","nSXCrcahhCNfzvXEY","ecatNAsyr4f2iQyGq","tpWeaaCgFjPTYCjg3"]}