Thomas Decomposition and Nonlinear Control Systems

Markus Lange‐Hegermann, Daniel RobertzarXiv 2020 · DOI · arXiv

TL;DR

This paper introduces the application of the Thomas decomposition technique, a symbolic method for analyzing systems of differential equations, to nonlinear control systems. By breaking down complex systems into simpler, "simple" systems, this approach allows for a more rigorous study of how system behavior and structural properties (like invertibility and observability) depend on parameters, offering a powerful tool for symbolic analysis in control theory.

What Problem Does This Paper Address?

Analyzing nonlinear control systems, especially their behavior under varying parameters, is a significant challenge. Traditional numerical methods can struggle with parameter dependence, while existing symbolic methods often lack a systematic way to handle the complexity and potential singularities that arise. The authors aim to bridge this gap by applying the Thomas decomposition, a technique from differential algebra, to provide a structured, symbolic framework for understanding the structural properties of nonlinear control systems and their dependence on parameters. This allows for a more in-depth analysis of concepts like invertibility, observability, and flat outputs, which are crucial for control system design and analysis.

Methods

The core of this paper is the application of the Thomas decomposition method. This technique, originally developed for systems of partial differential equations, involves symbolically decomposing a given system into a finite collection of "simple" systems. These simple systems are formally integrable and, when taken together, partition the solution set of the original system. The paper details both the algebraic and differential aspects of the Thomas decomposition.

Key aspects of the method include:

  • Algebraic Systems: The decomposition of systems of polynomial equations and inequations into simple algebraic systems, characterized by specific properties of their leading variables, initials, and discriminants.
  • Differential Systems: Extending the algebraic decomposition to systems involving differential equations. This involves concepts like rankings, Janet division, and the notion of "passive" and "formally integrable" systems.
  • Symbolic Computation: The entire process relies on symbolic manipulation, implemented using a Maple package. This allows for exact computations and explicit representation of parameter dependencies.
  • Control-Theoretic Applications: The paper demonstrates how the resulting simple systems can be used to analyze structural properties of control systems, such as invertibility, observability, and flatness, by examining the relationships between differential indeterminates (representing states, inputs, outputs, etc.) and their derivatives.

Datasets

Not discussed in the paper. The paper focuses on theoretical development and uses explicit examples derived from control theory problems rather than standard datasets.

Key Insights and Ideas

  • Parameter Dependence Unveiled: The Thomas decomposition provides a systematic way to reveal how structural properties of nonlinear control systems change with parameters. Different simple systems in the decomposition correspond to distinct structural behaviors, effectively partitioning the parameter space.
  • Bridging Differential Algebra and Control Theory: The paper successfully bridges the gap between advanced differential algebra techniques and practical problems in nonlinear control systems, offering a new symbolic toolset.
  • Formalizing Invertibility and Observability: The method allows for a rigorous symbolic definition and computation of invertibility, observability, and flat outputs, even in the presence of parameters.
  • Singular vs. Generic Cases: The decomposition naturally separates singular cases of system behavior from the generic case, which is crucial for robust control design.
  • Maple Implementation: The availability of a Maple package for Thomas decomposition makes these advanced techniques accessible for practical application and experimentation.

Key Results

The paper demonstrates the application of Thomas decomposition to several examples from control theory, including:

  • Unicycle Model: Inverting the unicycle model to express inputs in terms of outputs and their derivatives, revealing different configurations based on the decomposition.
  • 2-D Crane Model: Determining that the position coordinates {x, z} constitute a flat output for the crane model, showcasing the method's ability to identify flat outputs.
  • Continuous Stirred-Tank Reactor: Analyzing the dependence of the system's flatness on parameters (concentrations c1, c2), showing how parameter equality can preclude control.
  • Linear Partial Differential Equations: Studying the influence of a parametric function on infinitesimal transformations, highlighting how the decomposition reveals constraints on the parameter.

While no specific quantitative performance metrics (like speedup or accuracy improvements over other methods) are presented, the key result is the successful demonstration of the applicability and utility of Thomas decomposition for symbolic analysis of nonlinear control systems, particularly concerning parameter dependence and structural properties.

Limitations and Future Work

The authors acknowledge that a Thomas decomposition is not uniquely determined and depends on choices like the total ordering of variables and the order of intermediate steps. They also note that extending these techniques to nonlinear differential time-delay systems or nonlinear difference equations remains an open challenge, as analogs of Thomas decomposition are not yet known for these systems. Future work could explore these extensions and further applications of the method to more complex control problems.

Summary updated

Abstract

This paper applies the Thomas decomposition technique to nonlinear control systems, in particular to the study of the dependence of the system behavior on parameters. Thomas' algorithm is a symbolic method which splits a given system of nonlinear partial differential equations into a finite family of so-called simple systems which are formally integrable and define a partition of the solution set of the original differential system. Different simple systems of a Thomas decomposition describe different structural behavior of the control system in general. The paper gives an introduction to the Thomas decomposition method and shows how notions such as invertibility, observability and flat outputs can be studied. A Maple implementation of Thomas' algorithm is used to illustrate the techniques on explicit examples.
BibTeX
@misc{langehegermann2020thomas,
  author = {Markus Lange‐Hegermann and Daniel Robertz},
  title = {{Thomas Decomposition and Nonlinear Control Systems}},
  year = {2020},
  eprint = {2001.08424},
  archivePrefix = {arXiv},
  url = {https://arxiv.org/abs/2001.08424},
}

Loading related content...

Figure