Filters








8,340 Hits in 2.6 sec

Neural Lyapunov Model Predictive Control: Learning Safe Global Controllers from Sub-optimal Examples [article]

Mayank Mittal, Marco Gallieri, Alessio Quaglino, Seyed Sina Mirrazavi Salehian, Jan Koutník
2021 arXiv   pre-print
The terminal cost is constructed as a Lyapunov function neural network with the aim of recovering or extending the stable region of the initial demonstrator using a short prediction horizon.  ...  With a growing interest in data-driven control techniques, Model Predictive Control (MPC) provides an opportunity to exploit the surplus of data reliably, particularly while taking safety and stability  ...  Summary of contributions In this paper, we present Neural Lyapunov MPC, an algorithmic framework that obtains a single-step horizon Model Predictive Controller (MPC) for Lyapunov-based control of a non-linear  ... 
arXiv:2002.10451v2 fatcat:mpmbdjk6ozhsbkqptyl2zl55oq

Neural Lyapunov Differentiable Predictive Control [article]

Sayak Mukherjee, Ján Drgoňa, Aaron Tuor, Mahantesh Halappanavar, Draguna Vrabie
2022 arXiv   pre-print
The neural Lyapunov differentiable predictive control (NLDPC) learns the policy by constructing a computational graph encompassing the system dynamics, state and input constraints, and the necessary Lyapunov  ...  Our offline training approach provides a computationally efficient and scalable alternative to classical explicit model predictive control solutions.  ...  For the system dynamic model (1), we want to compute parametric predictive control policies modeled by deep neural networks π θ (x 0 ) : R nx → R N ×nu (4) that minimizes the parametric control objective  ... 
arXiv:2205.10728v1 fatcat:s3xhqomskrajje7ogqv7cea3wy

Rollover prediction and control in heavy vehicles via recurrent neural networks

E.N. Sanchez, L.J. Ricalde, R. Langari, D. Shahmirzadi
2004 2004 43rd IEEE Conference on Decision and Control (CDC) (IEEE Cat. No.04CH37601)  
In order to develop this control structure, a high order recurrent neural network is used to model the unknown tractor semitrailer system; a learning law is obtained using the Lyapunov methodology.  ...  Based on this prediction, an active control system is designed to prevent rollover.  ...  The neural identifier builds an on-line model for the trailer-semitrailer model which is assumed to be unknown. A learning adaptation law is derived using the Lyapunov methodology.  ... 
doi:10.1109/cdc.2004.1429635 fatcat:43jtr444bnc7pjuzlvdcieos6i

Stability analysis of embedded nonlinear predictor neural generalized predictive controller

Hesham F. Abdel Ghaffar, Sherif A. Hammad, Ahmed H. Yousef
2014 Alexandria Engineering Journal  
Nonlinear Predictor-Neural Generalized Predictive Controller (NGPC) is one of the most advanced control techniques that are used with severe nonlinear processes.  ...  Also, the superiority of NGPC over linear predictive controllers, like GPC, is proved for severe nonlinear processes over wide range of operation.  ...  Introduction Nowadays one of the most efficient networked controllers dealing with nonlinear process is the model predictive controllers' family.  ... 
doi:10.1016/j.aej.2013.11.008 fatcat:owa767sfdfc5penz7svozjxpji

LyaNet: A Lyapunov Framework for Training Neural ODEs [article]

Ivan Dario Jimenez Rodriguez, Aaron D. Ames, Yisong Yue
2022 arXiv   pre-print
We propose a method for training ordinary differential equations by using a control-theoretic Lyapunov condition for stability.  ...  Relative to standard Neural ODE training, we empirically find that LyaNet can offer improved prediction performance, faster convergence of inference dynamics, and improved adversarial robustness.  ...  ., 2020) , accurate time-series modeling (Chen et al., 2021a) , among others. Neural ODE Model Class.  ... 
arXiv:2202.02526v1 fatcat:aickbgmggbhozbk5cpggnospfm

Continuous Lyapunov Controller and Chaotic Non-linear System Optimization using Deep Machine Learning [article]

Amr Mahmoud, Youmna Ismaeil, Mohamed Zohdy
2020 arXiv   pre-print
The deep neural model predicts the parameter values that would best counteract the expected system in-stability.  ...  using deep machine learning regression model.  ...  Most recently, there is the introduction neural lyapunov control which proposes the use of deep learning to find the control and Lyapunov functions.  ... 
arXiv:2010.14746v2 fatcat:he5migngdbc4ff2ocgrexdbabm

Continuous Lyapunov Controlled Non-linear System Optimization Using Deep Learning with Memory

Amr Mahmoud, Youmna Ismaeil, Mohamed Zohdy
2020 International Journal of Control Science and Engineering  
The deep neural model predicts the parameter values that would best counteract the expected system in-stability.  ...  using deep machine learning regression model.  ...  Most recently, there is the introduction neural Lyapunov control which proposes the use of deep learning to find the control and Lyapunov functions.  ... 
doi:10.5923/j.control.20201002.01 fatcat:gscbzayx5zcmtkhfllpqw3dik4

Learning to Transfer Dynamic Models of Underactuated Soft Robotic Hands [article]

Liam Schramm, Avishai Sintov, Abdeslam Boularias
2020 arXiv   pre-print
We derive an upper bound on the Lyapunov exponent of a trained transition model, and demonstrate two approaches that make use of this insight.  ...  The most common way of doing this with neural networks is to take an existing neural network, and simply train it more with new data.  ...  models to improve accuracy in feed-forward control [22] .  ... 
arXiv:2005.10418v1 fatcat:cvbf6bivlfc7pkhgvbmrglzsei

Combined Prediction of Wind Power with Chaotic Time Series Analysis

Wang Qiang, Yang Yang
2014 Open Automation and Control Systems Journal  
The results show that the proposed model is more effective than single embedding dimension model and linear weighted combination model, and the prediction error of neural network combination is less than  ...  The combined model respectively makes use of linear weighted model and NNs method to achieve combination of several neural networks models through phase space reconstruction after wind power series chaotic  ...  Prediction Model Absolute Average Error % m = 7 7.11 m = 8 6.90 m = 9 6.92 Linear combination 6.86 Neural network combination 6.70 The Open Automation and Control Systems Journal  ... 
doi:10.2174/1874444301406010117 fatcat:f47rxsksd5eabgbgucjynb4nfe

The Mobile Robot Anti-disturbance vSLAM Navigation Algorithm based on RBF Neural Network

S.F. Wong, Z. Yu
2019 Procedia Manufacturing  
The algorithm is based on Lyapunov direct method controller with RBF neural network estimator.  ...  The algorithm is based on Lyapunov direct method controller with RBF neural network estimator.  ...  The advanced algorithm includes Deep Learning model for prediction, estimation and navigation.  ... 
doi:10.1016/j.promfg.2020.01.051 fatcat:pozouwnx5vepno3be3ynj46jfm

Networked Model Predictive Control Using a Wavelet Neural Network [article]

H. Khodabandehlou, M. Sami Fadali
2018 arXiv   pre-print
The model predictive controller (MPC) uses the model to predict the future outputs of the system over an extended prediction horizon and calculates the optimal future inputs by minimizing a controller  ...  In this study, we use a wavelet neural network with a feedforward component and a model predictive controller for online nonlinear system identification over a communication network.  ...  Lyapunov stability theory is used to prove the stability of the model predictive controller.  ... 
arXiv:1805.04549v1 fatcat:tzae42srmzbbjhhsv4poo3j2ia

Hopfield Neural Networks for Parametric Identification of Dynamical Systems

Miguel Atencia, Gonzalo Joya, Francisco Sandoval
2005 Neural Processing Letters  
Thus the neural method is validated as an on-line estimator of the time-varying parameters appearing in the model of a nonlinear physical system.  ...  In this work, a novel method, based upon Hopfield neural networks, is proposed for parameter estimation, in the context of system identification.  ...  Estimation with Hopfield and Tank Networks The neural network paradigm comprises a variety of computational models, among which Hopfield networks [8, 9] are feedback systems for which a Lyapunov function  ... 
doi:10.1007/s11063-004-3424-3 fatcat:qhosfuhohvespd26nvdfrbf6la

Rollover Prediction and Control in Heavy Vehicles Via Recurrent High Order Neural Networks

Edgar N. Sanchez, Luis J. Ricalde, Reza Langari, Danial Shahmirzadi
2011 Intelligent Automation and Soft Computing  
In order to develop this control structure, a high order recurrent neural network is used to model the unknown tractor semitrailer system; a learning law is obtained using the Lyapunov methodology.  ...  Based on this prediction, an active control system is designed to prevent rollover for these vehicles.  ...  An update law for the neural network weights is proposed via the Lyapunov methodology. The control law is synthesized using the Lyapunov methodology and the inverse optimal control approach.  ... 
doi:10.1080/10798587.2011.10643135 fatcat:ybkefr3urzhyjappmrnzck67cq

Quantum Lyapunov control with machine learning [article]

S. C. Hou, X. X. Yi
2018 arXiv   pre-print
Two designs are presented and illustrated where the feedforward neural network and the general regression neural network are used to select control schemes and design Lyapunov functions, respectively.  ...  Quantum state engineering is a central task in Lyapunov-based quantum control.  ...  At last, apply the trained neural network to predict control parameters for new initial states. Two designs are proposed to select control schemes and design Lyapunov functions.  ... 
arXiv:1808.02516v1 fatcat:h2rgft64greo3oubczc7zc4yyu

Applied Neural Intelligence to Modeling, Control, and Management of Human Systems and Environments

Toly Chen, P. Balasubramaniam, Quek Hiok Chai, Yi-Chi Wang
2012 Applied Computational Intelligence and Soft Computing  
Khan conducted a comparative study to predict the stock index values using soft computing models and a time series model.  ...  Ang developed the Lyapunov theory-based radial basis function neural network (RBFNN) for traffic sign recognition.  ...  Khan conducted a comparative study to predict the stock index values using soft computing models and a time series model.  ... 
doi:10.1155/2012/595041 fatcat:v4wqafgrybdvrj4t42oeum2cpi
« Previous Showing results 1 — 15 out of 8,340 results