Lie Group integrators for mechanical systems

Elena Celledoni, Ergys Çokaj, Andrea Leone, Davide Murari, Brynjulf Owren

Abstract

Since their introduction in the 1990s, Lie group integrators have become a method of choice for simulating multibody dynamics, shape analysis, and biophysical systems. This paper explores two primary classes of intrinsic Lie group integrators: Runge–Kutta–Munthe–Kaas (RKMK) methods and commutator-free Lie group integrators. We examine their application within Hamiltonian frameworks, particularly for problems on cotangent bundles of Lie groups. By leveraging variational principles, we derive symplectic Lie group integrators and introduce practical implementation techniques such as adaptive time-stepping. The efficacy of these methods is demonstrated through complex mechanical applications, including an N-fold spherical pendulum and a dual-rotor payload transport system.

1. Introduction

In computational physics and robotics, simulating the motion of rigid bodies requires navigating the curved geometry of rotation groups and manifolds. Standard numerical integration techniques, such as the classic Runge-Kutta family, are designed for Euclidean spaces ($\mathbb{R}^n$). When applied to systems constrained to manifolds—such as the rotation group $SO(3)$—these methods suffer from numerical drift. Over time, the computed state leaves the manifold, leading to non-physical results, such as rotation matrices that lose their orthogonality or unit length.

Lie group integrators offer a solution by being "intrinsic" to the manifold. By formulating the numerical steps directly within the Lie group structure, these methods ensure that the simulated state remains on the manifold to machine precision, preserving the geometric properties of the physical system without the need for ad-hoc projections or re-normalization.

2. Methodology: RKMK and Commutator-Free Methods

The research focuses on two robust families of Lie group integrators:

2.1 Runge–Kutta–Munthe–Kaas (RKMK) Methods

RKMK methods operate by "lifting" the differential equation from the manifold to its associated Lie algebra, which is a flat vector space. The integration is performed in the Lie algebra, and the result is mapped back to the Lie group using the exponential map. This approach transforms a non-linear problem on a manifold into a linear-like problem in a vector space, ensuring the manifold constraint is naturally satisfied.

2.2 Commutator-Free Lie Group Methods

While RKMK methods are powerful, they often require the calculation of "commutators"—mathematical operations that account for the non-commutative nature of rotations. These calculations can be computationally expensive. Commutator-free methods achieve high-order accuracy by using compositions of exponentials, avoiding the direct calculation of commutators and providing a more efficient alternative for complex multi-body systems.

3. Adaptive Time-Stepping

A significant contribution of this work is the implementation of adaptive time-stepping for Lie group methods. By utilizing "embedded pairs"—where a high-order step is compared against a lower-order step—the integrator can estimate local truncation error. This allows the algorithm to dynamically adjust the step size: tightening it during high-acceleration phases and expanding it during smoother motion, thereby optimizing computational resources without sacrificing accuracy.

4. Applications in Mechanics

The theoretical framework is validated through two sophisticated mechanical models:

4.1 The N-Fold Spherical Pendulum

The N-fold pendulum consists of a chain of $N$ rigid links connected by spherical joints. The authors model this by introducing the restriction of the adjoint action of the group $SE(3)$ (the Special Euclidean group) to the tangent bundle of the two-dimensional sphere, $TS^2$. This formulation allows the Lie group integrator to maintain the fixed length of each pendulum link automatically, demonstrating superior stability over standard solvers which exhibit significant phase-space drift.

4.2 Dual-Rotor Payload Transport

A more complex application involves the controlled path of a payload transported by two rotors. This system is modeled on a high-dimensional manifold: $$\mathbb{R}^6 \times \left(SO(3) \times \mathfrak{so}(3)\right)^2 \times (TS^2)^2$$ This model accounts for the position and velocity of the rotors, their orientations, and the spherical dynamics of the cables. Lie group integrators prove essential here, as the high dimensionality and coupled constraints make standard Euclidean integration highly unstable.

5. Results and Discussion

Numerical experiments confirm that the Lie group versions of 3rd and 4th-order Runge-Kutta methods maintain their theoretical convergence rates on curved manifolds. In the case of the 2-link pendulum, the adaptive controller successfully identified abrupt changes in angular acceleration (around $t \approx 2.2s$), adjusting the time step to maintain a specified error tolerance.

Furthermore, the preservation of geometric constraints—such as the unit length of pendulum arms—was maintained at the level of $10^{-15}$, effectively machine precision. This level of accuracy is critical for long-term simulations where cumulative errors in standard solvers would lead to a total breakdown of the physical model.

6. Conclusion

Lie group integrators provide a mathematically rigorous and computationally efficient framework for simulating mechanical systems on manifolds. By moving beyond Euclidean approximations, these methods preserve the fundamental geometry of motion. Future research aims to extend these intrinsic methods to infinite-dimensional systems, such as flexible Cosserat rods, which are vital for modeling medical instruments and advanced aerospace components.

Summary updated

Abstract

Since they were introduced in the 1990s, Lie group integrators have become a method of choice in many application areas. These include multibody dynamics, shape analysis, data science, image registration and biophysical simulations. Two important classes of intrinsic Lie group integrators are the Runge--Kutta--Munthe--Kaas methods and the commutator free Lie group integrators. We give a short introduction to these classes of methods. The Hamiltonian framework is attractive for many mechanical problems, and in particular we shall consider Lie group integrators for problems on cotangent bundles of Lie groups where a number of different formulations are possible. There is a natural symplectic structure on such manifolds and through variational principles one may derive symplectic Lie group integrators. We also consider the practical aspects of the implementation of Lie group integrators, such as adaptive time stepping. The theory is illustrated by applying the methods to two nontrivial applications in mechanics. One is the N-fold spherical pendulum where we introduce the restriction of the adjoint action of the group $SE(3)$ to $TS^2$, the tangent bundle of the two-dimensional sphere. Finally, we show how Lie group integrators can be applied to model the controlled path of a payload being transported by two rotors. This problem is modeled on $\mathbb{R}^6\times \left(SO(3)\times \mathfrak{so}(3)\right)^2\times (TS^2)^2$ and put in a format where Lie group integrators can be applied.
BibTeX
@article{celledoni2022lie,
  author = {Elena Celledoni and Ergys Çokaj and Andrea Leone and Davide Murari and Brynjulf Owren},
  title = {{Lie Group integrators for mechanical systems}},
  journal = {Int. J. Comput. Math.},
  year = {2022},
  doi = {10.1080/00207160.2021.1966772},
}
Figure