Learning of Causal Observable Functions for Koopman-DFL Lifting Linearization of Nonlinear Controlled Systems and Its Application to Excavation Automation

@article{SSelby2021LearningOC,
  title={Learning of Causal Observable Functions for Koopman-DFL Lifting Linearization of Nonlinear Controlled Systems and Its Application to Excavation Automation},
  author={Nicholas S. Selby and H. Harry Asada},
  journal={IEEE Robotics and Automation Letters},
  year={2021},
  volume={6},
  pages={6297-6304}
}
Effective and causal observable functions for low-order lifting linearization of nonlinear controlled systems are learned from data by using neural networks. While Koopman operator theory allows us to represent a nonlinear system as a linear system in an infinite-dimensional space of observables, exact linearization is guaranteed only for autonomous systems with no input, and finding effective observable functions for approximation with a low-order linear system remains an open question. Dual… 

Figures from this paper

Global, Unified Representation of Heterogenous Robot Dynamics Using Composition Operators

— The complexity of robot dynamics often pertains to the hybrid nature of dynamics, where governing dynamic equations consist of heterogenous equations that are switched depending on the state of the

Data-Driven Models for Control Engineering Applications Using the Koopman Operator

The investigated data-driven numerical approximation methods of the Koopman operator are suitable as a low-effort alternative for the design steps of model building and adaptation in the classical model-based controller design method.

Physics-BasedCausal LiftingLinearization ofNonlinear Control SystemsUnderpinnedby theKoopmanOperator ?

Methods for constructing causal linear models from nonlinear dynamical systems through lifting linearization underpinned by Koopman operator and physical system modeling theory are presented. Outputs

Physics-Based Causal Lifting Linearization of Nonlinear Control Systems Underpinned by the Koopman Operator

Methods for constructing causal linear models from nonlinear dynamical systems through lifting linearization underpinned by Koopman operator and physical system modeling theory are presented. Outputs

References

SHOWING 1-10 OF 46 REFERENCES

Active Learning of Dynamics for Data-Driven Control Using Koopman Operators

This paper presents an active learning strategy for robotic systems that takes into account task information, enables fast learning, and allows control to be readily synthesized by taking advantage

Dual Faceted Linearization of Nonlinear Dynamical Systems Based on Physical Modeling Theory

A new approach to modeling and linearization of nonlinear lumped-parameter systems based on physical modeling theory and a data-driven statistical method is presented and it is shown that the state equation of a nonlinear dynamical system having a bond graph model of integral causality is linear.

PyTorch: An Imperative Style, High-Performance Deep Learning Library

This paper details the principles that drove the implementation of PyTorch and how they are reflected in its architecture, and explains how the careful and pragmatic implementation of the key components of its runtime enables them to work together to achieve compelling performance.

Learning nonlinear state-space models using autoencoders

Continuous control of an underground loader using deep reinforcement learning

The reinforcement learning control of an underground loader was investigated in a simulated environment by using a multi-agent deep neural network approach, which learns policies for efficient bucket filling over many subsequent loading cycles, with a clear ability to adapt to the changing environment.

A multiscale model of terrain dynamics for real-time earthmoving simulation

A multiscale model for real-time simulation of terrain dynamics is explored, which is a computationally efficient model for earthmoving operations that resolve the motion of the soil, using a fast iterative solver, and provide realistic forces and dynamic for the equipment, uses a direct solver for high numerical precision.

Deep Learning of Koopman Representation for Control

A data-driven, model-free approach for the optimal control of the dynamical system that relies on the Deep Neural Network based learning of Koopman operator for the purpose of control.

MPC Performances for Nonlinear Systems Using Several Linearization Models

This paper compares performances of MPC (Model Predictive Control) for nonlinear systems using several different approximated linear models and it will be shown that computational times become much shorter and the optimality of the solutions are improved using particular lifting linearizations.

Dynamic Mode Decomposition and Robust Estimation: Case Study of a 2D Turbulent Boussinesq Flow

This paper focuses on an application of dynamic mode decomposition (DMD) identification methods and robust estimation theory to thermo-fluid systems modelled by the Boussinesq equations. First, we

Time Variable Minimum Torque Trajectory Optimization for Autonomous Excavator

The method formulates the excavation motion generation as a trajectory optimization problem and takes into account geometric, kinematic and dynamics constraints, and can generate feasible trajectories, which satisfy excavation task constraints and are adaptive to different soil conditions.