Sparse identification of nonlinear dynamics for model predictive control in the low-data limit

DOI10.1098/RSPA.2018.0335zbMATH Open1425.93175arXiv1711.05501OpenAlexW3100769368WikidataQ59807745 ScholiaQ59807745MaRDI QIDQ5243602FDOQ5243602

Authors: Eurika Kaiser, S. L. Brunton, J. N. Kutz 

Publication date: 19 November 2019

Published in: Proceedings of the Royal Society A: Mathematical, Physical and Engineering Sciences (Search for Journal in Brave)

Abstract: The data-driven discovery of dynamics via machine learning is currently pushing the frontiers of modeling and control efforts, and it provides a tremendous opportunity to extend the reach of model predictive control. However, many leading methods in machine learning, such as neural networks, require large volumes of training data, may not be interpretable, do not easily include known constraints and symmetries, and often do not generalize beyond the attractor where models are trained. These factors limit the use of these techniques for the online identification of a model in the low-data limit, for example following an abrupt change to the system dynamics. In this work, we extend the recent sparse identification of nonlinear dynamics (SINDY) modeling procedure to include the effects of actuation and demonstrate the ability of these models to enhance the performance of model predictive control (MPC), based on limited, noisy data. SINDY models are parsimonious, identifying the fewest terms in the model needed to explain the data, making them interpretable, generalizable, and reducing the burden of training data. We show that the resulting SINDY-MPC framework has higher performance, requires significantly less data, and is more computationally efficient and robust to noise than neural network models, making it viable for online training and execution in response to rapid changes to the system. SINDY-MPC also shows improved performance over linear data-driven models, although linear models may provide a stopgap until enough data is available for SINDY. SINDY-MPC is demonstrated on a variety of dynamical systems with different challenges, including the chaotic Lorenz system, a simple model for flight control of an F8 aircraft, and an HIV model incorporating drug treatment.

Full work available at URL: https://arxiv.org/abs/1711.05501

Recommendations

• SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
• Data-driven sparse identification of nonlinear dynamical systems using linear multistep methods
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Sparse identification of nonlinear dynamics with low-dimensionalized flow representations
• Weak SINDy: Galerkin-based data-driven model selection

zbMATH Keywords

machine learningsystem identificationcontrol theorymodel predictive controlnonlinear dynamicssparse identification of nonlineardynamics

Mathematics Subject Classification ID

Learning and adaptive systems in artificial intelligence (68T05) Discrete-time control/observation systems (93C55)

Cites Work

• Constrained sparse Galerkin regression
• Nonlinear total variation based noise removal algorithms
• Title not available (Why is that?)
• Model predictive control: Theory and practice - a survey
• Just relax: convex programming methods for identifying sparse signals in noise
• Discovering governing equations from data by sparse identification of nonlinear dynamical systems
• Fitting autoregressive models for prediction
• Compressive sensing and low-rank libraries for classification of bifurcation regimes in nonlinear dynamical systems
• Dynamic mode decomposition of numerical and experimental data
• Spectral analysis of nonlinear flows
• Automated reverse engineering of nonlinear dynamical systems
• Numerical differentiation of noisy, nonsmooth data
• Deterministic Nonperiodic Flow
• Turbulence, coherent structures, dynamical systems and symmetry.
• System Identification Via Sparse Multiple Kernel-Based Regularization Using Sequential Convex Optimization Techniques
• Title not available (Why is that?)
• Mathematical Analysis of HIV-1 Dynamics in Vivo
• Kernel methods in system identification, machine learning and function estimation: a survey
• System identification in the presence of outliers and random noises: a compressed sensing approach
• Multilayer feedforward networks are universal approximators
• Title not available (Why is that?)
• The influence of diseases on Lotka-Volterra systems
• Model predictive control: recent developments and future promise
• Spectral properties of dynamical systems, model reduction and decompositions
• On dynamic mode decomposition: theory and applications
• Dynamic mode decomposition with control
• Applied Koopmanism
• An eigensystem realization algorithm for modal parameter identification and model reduction
• A data-driven approximation of the koopman operator: extending dynamic mode decomposition
• A model predictive control based scheduling method for HIV therapy
• Nonlinear system identification. NARMAX methods in the time, frequency, and spatio-temporal domains
• Title not available (Why is that?)
• Title not available (Why is that?)
• Design of nonlinear automatic flight control systems
• Analytical Foundations of Volterra Series
• Global optimal feedback control for general nonlinear systems with nonquadratic performance criteria
• Volterra series and geometric control theory
• A fast nonlinear model identification method
• Sparse identification of posynomial models
• Identifiability of affine linear parameter-varying models
• A Sparse Bayesian Approach to the Identification of Nonlinear State-Space Systems
• Sparse reduced-order modelling: sensor-based dynamics to full-state estimation
• Model selection for dynamical systems via sparse regression and information criteria
• Learning partial differential equations via data discovery and sparse optimization

Cited In (69)

• Recovery guarantees for polynomial coefficients from weakly dependent data with outliers
• Supervised learning algorithms for controlling underactuated dynamical systems
• IDENT: identifying differential equations with numerical time evolution
• On the convergence of the SINDy algorithm
• Incorporating physical constraints in a deep probabilistic machine learning framework for coarse-graining dynamical systems
• A Bayesian framework for learning governing partial differential equation from data
• On the universal transformation of data-driven models to control systems
• Discovery of Nonlinear Multiscale Systems: Sampling Strategies and Embeddings
• Sparse reduced-order modelling: sensor-based dynamics to full-state estimation
• SubTSBR to tackle high noise and outliers for data-driven discovery of differential equations
• The \(l_1\)-based sparsification of energy interactions in unsteady lid-driven cavity flow
• Data-driven identification of parametric partial differential equations
• Deep reinforcement learning for adaptive mesh refinement
• Detection and prediction of equilibrium states in kinetic plasma simulations via mode tracking using reduced-order dynamic mode decomposition
• Memory-based reduced modelling and data-based estimation of opinion spreading
• Local parameter identification with neural ordinary differential equations
• On Bilinear Time-Domain Identification and Reduction in the Loewner Framework
• Learning equations from biological data with limited time samples
• Conditional space-time POD extensions for stability and prediction analysis
• Model selection via reweighted partial sparse recovery
• Hybrid analysis and modeling, eclecticism, and multifidelity computing toward digital twin revolution
• On the structure of time-delay embedding in linear models of non-linear dynamical systems
• Sparse identification of nonlinear dynamics with low-dimensionalized flow representations
• On the differential realization of a second-order bilinear system in a Hilbert space
• Group projected subspace pursuit for identification of variable coefficient differential equations (GP-IDENT)
• SINDy-PI: a robust algorithm for parallel implicit sparse identification of nonlinear dynamics
• Poincaré maps for multiscale physics discovery and nonlinear Floquet theory
• From the POD-Galerkin method to sparse manifold models
• Deep learning‐based reduced order models for the real‐time simulation of the nonlinear dynamics of microstructures
• Simultaneous Identification and Denoising of Dynamical Systems
• A framework based on symbolic regression coupled with eXtended physics-informed neural networks for gray-box learning of equations of motion from data
• Reduced order modeling of parametrized systems through autoencoders and SINDy approach: continuation of periodic solutions
• Robust Identification of Differential Equations by Numerical Techniques from a Single Set of Noisy Observation
• Bounded nonlinear forecasts of partially observed geophysical systems with physics-constrained deep learning
• Phase-based control of periodic flows
• Data‐enabled predictive control for quadcopters
• Koopman Linear Quadratic Regulator Using Complex Eigenfunctions for Nonlinear Dynamical Systems
• DL-PDE: Deep-Learning Based Data-Driven Discovery of Partial Differential Equations from Discrete and Noisy Data
• Learning sparse nonlinear dynamics via mixed-integer optimization
• Deep learning discrete calculus (DLDC): a family of discrete numerical methods by universal approximation for STEM education to frontier research
• Data-driven approximation of the Koopman generator: model reduction, system identification, and control
• Proximal policy optimization‐based controller for chaotic systems
• Probabilistic forecast of nonlinear dynamical systems with uncertainty quantification
• Some applications of compressed sensing in computational mechanics: model order reduction, manifold learning, data-driven applications and nonlinear dimensionality reduction
• Thermodynamically consistent physics-informed neural networks for hyperbolic systems
• Discovering stochastic partial differential equations from limited data using variational Bayes inference
• Kernel functions embed into the autoencoder to identify the sparse models of nonlinear dynamics
• Sparse identification of dynamical systems by reweighted \(l_1\)-regularized least absolute deviation regression
• Neural network-based parametric system identification: a review
• Learning-based model predictive current control for synchronous machines: an LSTM approach
• Modern Koopman theory for dynamical systems
• On the sample complexity of stabilizing linear dynamical systems from data
• Structured model selection via ℓ1−ℓ2 optimization
• On higher order drift and diffusion estimates for stochastic SINDy
• Distributed sparse identification for stochastic dynamic systems under cooperative non-persistent excitation condition
• SPADE4: sparsity and delay embedding based forecasting of epidemics
• A thermodynamics-informed neural network for elastoplastic constitutive modeling of granular materials
• Dynamical system identification, model selection, and model uncertainty quantification by Bayesian inference
• Online identification and control of PDEs via reinforcement learning methods
• EKF-SINDy: empowering the extended Kalman filter with sparse identification of nonlinear dynamics
• Data-driven transient lift attenuation for extreme vortex gust-airfoil interactions
• Deep learning discovery of macroscopic governing equations for viscous gravity currents from microscopic simulation data
• A stochastic gradient-based two-step sparse identification algorithm for multivariate ARX systems
• Forecasting and predicting stochastic agent-based model data with biologically-informed neural networks
• Self-tuning model predictive control for wake flows
• KAN-ODEs: Kolmogorov-Arnold network ordinary differential equations for learning dynamical systems and hidden physics
• Reinforcement learning based MPC with neural dynamical models
• Dictionary-based model reduction for state estimation
• Model predictive complex system control from observational and interventional data

Uses Software

• SINDy

This page was built for publication: Sparse identification of nonlinear dynamics for model predictive control in the low-data limit

Report a bug (only for logged in users!)Click here to report a bug for this page (MaRDI item Q5243602)