-
Neural networks for the approximation of Euler's elastica
Authors:
Elena Celledoni,
Ergys Çokaj,
Andrea Leone,
Sigrid Leyendecker,
Davide Murari,
Brynjulf Owren,
Rodrigo T. Sato Martín de Almagro,
Martina Stavole
Abstract:
Euler's elastica is a classical model of flexible slender structures, relevant in many industrial applications. Static equilibrium equations can be derived via a variational principle. The accurate approximation of solutions of this problem can be challenging due to nonlinearity and constraints. We here present two neural network based approaches for the simulation of this Euler's elastica. Starti…
▽ More
Euler's elastica is a classical model of flexible slender structures, relevant in many industrial applications. Static equilibrium equations can be derived via a variational principle. The accurate approximation of solutions of this problem can be challenging due to nonlinearity and constraints. We here present two neural network based approaches for the simulation of this Euler's elastica. Starting from a data set of solutions of the discretised static equilibria, we train the neural networks to produce solutions for unseen boundary conditions. We present a $\textit{discrete}$ approach learning discrete solutions from the discrete data. We then consider a $\textit{continuous}$ approach using the same training data set, but learning continuous solutions to the problem. We present numerical evidence that the proposed neural networks can effectively approximate configurations of the planar Euler's elastica for a range of different boundary conditions.
△ Less
Submitted 4 June, 2024; v1 submitted 1 December, 2023;
originally announced December 2023.
-
Discrete Adjoint Method for Variational Integration of Constrained ODEs and its application to Optimal Control of Geometrically Exact Beam Dynamics
Authors:
Matthias Schubert,
Rodrigo T. Sato Martín de Almagro,
Karin Nachbagauer,
Sina Ober-Blöbaum,
Sigrid Leyendecker
Abstract:
Direct methods for the simulation of optimal control problems apply a specific discretization to the dynamics of the problem, and the discrete adjoint method is suitable to calculate corresponding conditions to approximate an optimal solution. While the benefits of structure preserving or geometric methods have been known for decades, their exploration in the context of optimal control problems is…
▽ More
Direct methods for the simulation of optimal control problems apply a specific discretization to the dynamics of the problem, and the discrete adjoint method is suitable to calculate corresponding conditions to approximate an optimal solution. While the benefits of structure preserving or geometric methods have been known for decades, their exploration in the context of optimal control problems is a relatively recent field of research. In this work, the discrete adjoint method is derived for variational integrators yielding structure preserving approximations of the dynamics firstly in the ODE case and secondly for the case in which the dynamics is subject to holonomic constraints. The convergence rates are illustrated by numerical examples. Thirdly, the discrete adjoint method is applied to geometrically exact beam dynamics, represented by a holonomically constrained PDE.
△ Less
Submitted 27 November, 2023;
originally announced November 2023.
-
A new Lagrangian approach to control affine systems with a quadratic Lagrange term
Authors:
Sigrid Leyendecker,
Sofya Maslovskaya,
Sina Ober-Blobaum,
Rodrigo T. Sato Martin de Almagro,
Flora Orsolya Szemenyei
Abstract:
In this work, we consider optimal control problems for mechanical systems on vector spaces with fixed initial and free final state and a quadratic Lagrange term. Specifically, the dynamics is described by a second order ODE containing an affine control term and we allow linear coordinate changes in the configuration space. Classically, Pontryagin's maximum principle gives necessary optimality cond…
▽ More
In this work, we consider optimal control problems for mechanical systems on vector spaces with fixed initial and free final state and a quadratic Lagrange term. Specifically, the dynamics is described by a second order ODE containing an affine control term and we allow linear coordinate changes in the configuration space. Classically, Pontryagin's maximum principle gives necessary optimality conditions for the optimal control problem. For smooth problems, alternatively, a variational approach based on an augmented objective can be followed. Here, we propose a new Lagrangian approach leading to equivalent necessary optimality conditions in the form of Euler-Lagrange equations. Thus, the differential geometric structure (similar to classical Lagrangian dynamics) can be exploited in the framework of optimal control problems. In particular, the formulation enables the symplectic discretisation of the optimal control problem via variational integrators in a straightforward way.
△ Less
Submitted 26 February, 2024; v1 submitted 25 July, 2023;
originally announced July 2023.
-
A new perspective on symplectic integration of constrained mechanical systems via discretization maps
Authors:
María Barbero Liñán,
David Martín de Diego,
Rodrigo T. Sato Martín de Almagro
Abstract:
A new procedure to construct symplectic methods for constrained mechanical systems is developed in this paper. The definition of a map coming from the notion of retraction maps allows to adapt the continuous problem to the discretization rule rather than viceversa. As a result, the constraint submanifold is exactly preserved by the discrete flow and the extension of the methods to the case of non-…
▽ More
A new procedure to construct symplectic methods for constrained mechanical systems is developed in this paper. The definition of a map coming from the notion of retraction maps allows to adapt the continuous problem to the discretization rule rather than viceversa. As a result, the constraint submanifold is exactly preserved by the discrete flow and the extension of the methods to the case of non-linear configuration spaces is straightforward.
△ Less
Submitted 11 June, 2023;
originally announced June 2023.
-
A parallel iterative method for variational integration
Authors:
Sebastián J. Ferraro,
David Martín de Diego,
Rodrigo Takuro Sato Martín de Almagro
Abstract:
Discrete variational methods show excellent performance in numerical simulations of different mechanical systems. In this paper, we introduce an iterative procedure for the solution of discrete variational equations for boundary value problems. More concretely, we explore a parallelization strategy that leverages the capabilities of multicore CPUs and GPUs (graphics cards). We study this parallel…
▽ More
Discrete variational methods show excellent performance in numerical simulations of different mechanical systems. In this paper, we introduce an iterative procedure for the solution of discrete variational equations for boundary value problems. More concretely, we explore a parallelization strategy that leverages the capabilities of multicore CPUs and GPUs (graphics cards). We study this parallel method for higher-order Lagrangian systems, which appear in fully-actuated problems and beyond. The most important part of the paper is devoted to a precise study of different convergence conditions for these methods. We illustrate their excellent behavior in some interesting examples, namely Zermelo's navigation problem, a fuel-optimal navigation problem, interpolation problems or in a fuel optimization problem for a controlled 4-body problem in astrodynamics showing the potential of our method.
△ Less
Submitted 17 June, 2022;
originally announced June 2022.
-
High-order integrators for Lagrangian systems on homogeneous spaces via nonholonomic mechanics
Authors:
Rodrigo T. Sato Martín de Almagro
Abstract:
In this paper, high-order numerical integrators on homogeneous spaces will be presented as an application of nonholonomic partitioned Runge-Kutta Munthe-Kaas (RKMK) methods on Lie groups.
A homogeneous space $M$ is a manifold where a group $G$ acts transitively. Such a space can be understood as a quotient $M \cong G/H$, where $H$ a closed Lie subgroup, is the isotropy group of each point of…
▽ More
In this paper, high-order numerical integrators on homogeneous spaces will be presented as an application of nonholonomic partitioned Runge-Kutta Munthe-Kaas (RKMK) methods on Lie groups.
A homogeneous space $M$ is a manifold where a group $G$ acts transitively. Such a space can be understood as a quotient $M \cong G/H$, where $H$ a closed Lie subgroup, is the isotropy group of each point of $M$. The Lie algebra of $G$ may be decomposed into $\mathfrak{g} = \mathfrak{m} \oplus \mathfrak{h}$, where $\mathfrak{h}$ is the subalgebra that generates $H$ and $\mathfrak{m}$ is a subspace. Thus, variational problems on $M$ can be treated as nonholonomically constrained problems on $G$, by requiring variations to remain on $\mathfrak{m}$.
Nonholonomic partitioned RKMK integrators are derived as a modification of those obtained by a discrete variational principle on Lie groups, and can be interpreted as obeying a discrete Chetaev principle. These integrators tend to preserve several properties of their purely variational counterparts.
△ Less
Submitted 28 January, 2022;
originally announced January 2022.
-
Parallel iterative methods for variational integration applied to navigation problems
Authors:
Sebastián J. Ferraro,
David Martín de Diego,
Rodrigo T. Sato Martín de Almagro
Abstract:
Discrete variational methods have shown an excellent performance in numerical simulations of different mechanical systems. In this paper, we introduce an iterative method for discrete variational methods appropriate for boundary value problems. More concretely, we explore a parallelization strategy that leverages the power of multicore CPUs and GPUs (graphics cards). We study this parallel method…
▽ More
Discrete variational methods have shown an excellent performance in numerical simulations of different mechanical systems. In this paper, we introduce an iterative method for discrete variational methods appropriate for boundary value problems. More concretely, we explore a parallelization strategy that leverages the power of multicore CPUs and GPUs (graphics cards). We study this parallel method for first-order and second-order Lagrangians and we illustrate its excellent behavior in some interesting applications, namely Zermelo's navigation problem, a fuel-optimal navigation problem, and an interpolation problem.
△ Less
Submitted 12 September, 2021;
originally announced September 2021.
-
Geometric Optimal Trajectory Tracking of Nonholonomic Mechanical Systems
Authors:
Leonardo J. Colombo,
David Martín de Diego,
Aradhana Nayak,
Rodrigo T. Sato Martín de Almagro
Abstract:
We study the tracking of a trajectory for a nonholonomic system by recasting the problem as a constrained optimal control problem. The cost function is chosen to minimize the error in positions and velocities between the trajectory of a nonholonomic system and the desired reference trajectory, both evolving on the distribution which defines the nonholonomic constraints. The problem is studied from…
▽ More
We study the tracking of a trajectory for a nonholonomic system by recasting the problem as a constrained optimal control problem. The cost function is chosen to minimize the error in positions and velocities between the trajectory of a nonholonomic system and the desired reference trajectory, both evolving on the distribution which defines the nonholonomic constraints. The problem is studied from a geometric framework. Optimality conditions are determined by the Pontryagin Maximum Principle and also from a variational point of view, which allows the construction of geometric integrators. Examples and numerical simulations are shown to validate the results.
△ Less
Submitted 25 January, 2020;
originally announced January 2020.
-
Variational order for forced Lagrangian systems II: Euler-Poincaré equations with forcing
Authors:
David Martín de Diego,
Rodrigo T. Sato Martín de Almagro
Abstract:
In this paper we provide a variational derivation of the Euler-Poincaré equations for systems subjected to external forces using an adaptation of the techniques introduced by Galley and others. Moreover, we study in detail the underlying geometry which is related to the notion of Poisson groupoid. Finally, we apply the previous construction to the formal derivation of the variational error for num…
▽ More
In this paper we provide a variational derivation of the Euler-Poincaré equations for systems subjected to external forces using an adaptation of the techniques introduced by Galley and others. Moreover, we study in detail the underlying geometry which is related to the notion of Poisson groupoid. Finally, we apply the previous construction to the formal derivation of the variational error for numerical integrators of forced Euler-Poincaré equations and the application of this theory to the derivation of geometric integrators for forced systems.
△ Less
Submitted 24 June, 2019;
originally announced June 2019.
-
High-order geometric methods for nonholonomic mechanical systems
Authors:
David Martín de Diego,
Rodrigo T. Sato Martín de Almagro
Abstract:
In the last two decades, significant effort has been put in understanding and designing so-called structure-preserving numerical methods for the simulation of mechanical systems. Geometric integrators attempt to preserve the geometry associated to the original system as much as possible, such as the structure of the configuration space, the energy behaviour, preservation of constants of the motion…
▽ More
In the last two decades, significant effort has been put in understanding and designing so-called structure-preserving numerical methods for the simulation of mechanical systems. Geometric integrators attempt to preserve the geometry associated to the original system as much as possible, such as the structure of the configuration space, the energy behaviour, preservation of constants of the motion and of constraints or other structures associated to the continuous system (symplecticity, Poisson structure...). In this article, we develop high-order geometric (or pseudo-variational) integrators for nonholonomic systems, i.e., mechanical systems subjected to constraint functions which are, roughly speaking, functions on velocities that are not derivable from position constraints. These systems realize rolling or certain kinds of sliding contact and are important for describing different classes of vehicles.
△ Less
Submitted 25 October, 2018;
originally announced October 2018.
-
Convergence of Lobatto-type Runge-Kutta methods for partitioned differential-algebraic systems of index 2
Authors:
Rodrigo Takuro Sato Martín de Almagro
Abstract:
In this paper we propose a numerical scheme for partitioned systems of index 2 DAEs, such as those arising from nonholonomic mechanical problems and prove the order of a certain class of Runge-Kutta methods we call of Lobatto-type. The study of nonholonomic systems has recently shown a new interest in that theory and also in its relation to the new developments in control theory, subriemannian geo…
▽ More
In this paper we propose a numerical scheme for partitioned systems of index 2 DAEs, such as those arising from nonholonomic mechanical problems and prove the order of a certain class of Runge-Kutta methods we call of Lobatto-type. The study of nonholonomic systems has recently shown a new interest in that theory and also in its relation to the new developments in control theory, subriemannian geometry, robotics, etc. The proofs and general outline of the paper follow a similar procedure of the one by L.O. Jay in the non-partitioned setting, but we tackle the issue of having two different sets of coefficients in interaction.
△ Less
Submitted 24 October, 2018;
originally announced October 2018.
-
The missing light of the Hubble Ultra Deep Field
Authors:
Alejandro Borlaff,
Ignacio Trujillo,
Javier Román,
John E. Beckman,
M. Carmen Eliche-Moral,
Raúl Infante-Sáinz,
Alejandro Lumbreras,
Rodrigo Takuro Sato Martín de Almagro,
Carlos Gómez-Guijarro,
María Cebrián,
Antonio Dorta,
Nicolás Cardiel,
Mohammad Akhlaghi,
Cristina Martínez-Lombilla
Abstract:
The Hubble Ultra Deep field (HUDF) is the deepest region ever observed with the Hubble Space Telescope. With the main objective of unveiling the nature of galaxies up to $z \sim 7-8$, the observing and reduction strategy have focused on the properties of small and unresolved objects, rather than the outskirts of the largest objects, which are usually over-subtracted.
We aim to create a new set o…
▽ More
The Hubble Ultra Deep field (HUDF) is the deepest region ever observed with the Hubble Space Telescope. With the main objective of unveiling the nature of galaxies up to $z \sim 7-8$, the observing and reduction strategy have focused on the properties of small and unresolved objects, rather than the outskirts of the largest objects, which are usually over-subtracted.
We aim to create a new set of WFC3/IR mosaics of the HUDF using novel techniques to preserve the properties of the low surface brightness regions. We created ABYSS: a pipeline that optimises the estimate and modelling of low-level systematic effects to obtain a robust background subtraction.
We have improved four key points in the reduction: 1) creation of new absolute sky flat fields, 2) extended persistence models, 3) dedicated sky background subtraction and 4) robust co-adding. The new mosaics successfully recover the low surface brightness structure removed on the previous HUDF published reductions.
The amount of light recovered with a mean surface brightness dimmer than $\overlineμ=26$ mar arcsec$^{-2}$ is equivalent to a m=19 mag source when compared to the XDF and a m=20 mag compared to the HUDF12. We present a set of techniques to reduce ultra-deep images ($μ>32.5$ mag arcsec$^{-2}$, $3σ$ in $10\times10$ arcsec boxes), that successfully allow to detect the low surface brightness structure of extended sources on ultra deep surveys. The developed procedures are applicable to HST, JWST, EUCLID and many other space and ground-based observatories. We will make the final ABYSS WFC3/IR HUDF mosaics publicly available at http://www.iac.es/proyecto/abyss/.
△ Less
Submitted 4 February, 2019; v1 submitted 28 September, 2018;
originally announced October 2018.
