Global controllability tests for geometric hybrid control systems

María Barbero-Liñán, Jorge Cortés, David Martı́n de Diego, Sonia Martı́nez, Miguel C. Muñoz‐Lecanda

TL;DR

This paper introduces a novel geometric framework for hybrid control systems—systems that blend continuous motion with discrete "jumps." The authors demonstrate that the geometry of the jump set itself can be a powerful tool for control, proving that a hybrid system can be globally controllable even when none of its individual continuous components are.

What Problem Does This Paper Address?

In control theory, "controllability" asks a fundamental question: can you get from point A to point B using the available inputs? For standard continuous systems (like a car or a drone), we have well-established tools to answer this. However, many real-world systems are hybrid: a car shifting gears, a walking robot hitting the ground, or a thermostat clicking on and off.

The challenge is that hybrid systems are notoriously difficult to analyze. Previous research often relied on "strong assumptions," such as assuming you can switch between different dynamics at any point in space. In reality, jumps usually happen only under specific conditions (the "jump set"). This paper fills the gap by providing a rigorous geometric language to describe these systems and developing tests to determine if they are globally controllable by looking at how the continuous trajectories interact with the geometry of the jump sets.

Methods

The authors build their framework using tools from differential geometry:

  • Generalized Dynamical Systems (GDS): They define the "building blocks" of hybrid systems as fiber bundles equipped with differentiable maps and submanifolds that represent constraints.
  • Geometric Hybrid Systems: They model the entire system as a directed graph where nodes are different dynamics (GDS) and edges are transitions. Crucially, they define a jump map ($R$) that dictates how the state changes during a transition.
  • Orbit Theorem & Foliations: For systems that are linear or affine in their controls, the authors use the Orbit Theorem to describe the "leaves" (the reachable sets) of the continuous dynamics.
  • Transversality Conditions: The core of their new controllability tests involves checking if the "leaves" of the continuous dynamics intersect the jump sets in a way that allows the system to "hop" from one reachable region to another.

Datasets

Not discussed in the paper. As a theoretical work in mathematical control theory, the paper relies on analytical proofs and illustrative synthetic examples rather than empirical datasets.

Key Insights and Ideas

  • Jumps as a Resource: Traditionally, discrete jumps are seen as complications. This paper flips the script, showing that the jump map can actually "gain" controllability. By carefully choosing where and how a system jumps, you can bridge gaps between regions that are otherwise unreachable.
  • The Power of Set-Valued Maps: The authors highlight that if a jump map is "set-valued" (meaning one jump point can lead to multiple possible successor points), it significantly boosts the system's ability to reach new states.
  • Geometry Matters: Controllability isn't just about the equations of motion; it's about the topology of the jump set. If the tangent space of the jump set and the tangent space of the dynamical "leaves" together span the entire configuration space, the system can move in directions that neither the jump nor the continuous motion could achieve alone.
  • Discrete vs. Global Controllability: The paper distinguishes between being able to switch between all "modes" (discrete controllability) and being able to reach any physical state (global controllability). You need both for a functional hybrid system.

Key Results

  • Theorem 4.6: Provides a necessary and sufficient algorithmic condition for global controllability based on sequences of discrete modes and jump points.
  • Theorem 4.13: Offers a more practical test. If a hybrid system has a strongly connected graph and its continuous "leaves" intersect the jump sets such that their combined tangent spaces cover the manifold, the system is globally controllable.
  • The "Submarine" Example: The authors present a compelling example of a system that can only move horizontally at certain depths (like a submarine restricted by tides). Individually, these horizontal modes are not controllable. However, by adding "instantaneous" vertical jumps at specific locations, the entire 2D space becomes reachable.

Limitations and Future Work

  • The "Dead End" Problem: The authors acknowledge that trajectories can sometimes hit a "dead end" where inequality constraints are no longer satisfied after a jump, preventing further motion.
  • Geometric Integration: A major hurdle for the future is "hybrid time discretization." When you turn these continuous geometric models into computer simulations, it is difficult to ensure the state hits the jump set exactly at the right time/position.
  • Optimization: While this paper focuses on whether you can reach a point, it doesn't yet address the best way to get there (optimal control). The authors suggest their geometric framework is the ideal foundation for extending classical optimal control results to the hybrid world.

Summary updated

Abstract

Hybrid systems are characterized by having an interaction between continuous dynamics and discrete events. The contribution of this paper is to provide hybrid systems with a novel geometric formulation so that controls can be added. Using this framework we describe some new global controllability tests for hybrid control systems exploiting the geometry and the topology of the set of jump points, where the instantaneous change of dynamics take place. Controllability is understood as the existence of a feasible trajectory for the system joining any two given points. As a result we describe examples where none of the continuous control systems are controllable, but the associated hybrid system is controllable because of the characteristics of the jump set.
BibTeX
@inproceedings{barberolinan2020global,
  author = {María Barbero-Liñán and Jorge Cortés and David Martı́n de Diego and Sonia Martı́nez and Miguel C. Muñoz‐Lecanda},
  title = {{Global controllability tests for geometric hybrid control systems}},
  booktitle = {Nonlinear},
  year = {2020},
  doi = {10.1016/j.nahs.2020.100935},
}

Loading related content...

Figure