Robot-assisted technologies are increasingly being developed for their use in endovascular surgery. These technologies can aid clinicians in providing higher precision, increased accuracy, and reproducibility. This paper further extends the clinical feasibility of an autonomous system designed to provide assistance during endovascular procedures. We develop a control strategy that allows a user to guide a magnetic catheter to target sites. The system handles the full synchronization of co-operating manipulators that both actuate a clinical tool. The catheter is positioned using magnetic actuation and proportional-integral (PI) control in conjunction with real-time ultrasound images. Our results indicate an average error between the tracked catheter tip and target positions of 2.09±0.49 mm. We expect that our system will provide a step towards collaborative manipulators employing mobile electromagnets, and possibly improve autonomous catheterization procedures within endovascular surgeries.
22 June 2023 in Optimal Control Applications and Methods
It is well known that conservative mechanical systems exhibit local oscillatory behaviors due to their elastic and gravitational potentials, which completely characterize these periodic motions together with the inertial properties of the system. The classification of these periodic behaviors and their geometric characterization are in an ongoing secular debate, which recently led to the so‐called eigenmanifold theory. The eigenmanifold characterizes nonlinear oscillations as a generalization of linear eigenspaces. With the motivation of performing periodic tasks efficiently, we use tools coming from this theory to construct an optimization problem aimed at inducing desired closed‐loop oscillations through a state feedback law. We solve the constructed optimization problem via gradient‐descent methods involving neural networks. Extensive simulations show the validity of the approach.
14 June 2024 in The International Journal of Robotics Research
This work presents a novel approach for the optimization of dynamic systems on finite-dimensional Lie groups. We rephrase dynamic systems as so-called neural ordinary differential equations (neural ODEs), and formulate the optimization problem on Lie groups. A gradient descent optimization algorithm is presented to tackle the optimization numerically. Our algorithm is scalable, and applicable to any finite-dimensional Lie group, including matrix Lie groups. By representing the system at the Lie algebra level, we reduce the computational cost of the gradient computation. In an extensive example, optimal potential energy shaping for control of a rigid body is treated. The optimal control problem is phrased as an optimization of a neural ODE on the Lie group SE(3), and the controller is iteratively optimized. The final controller is validated on a state-regulation task.
Neural ordinary differential equations (neural ODEs) are a well-established tool for optimizing the parameters of dynamical systems, with applications in image classification, optimal control, and physics learning. Although dynamical systems of interest often evolve on Lie groups and more general differentiable manifolds, theoretical results for neural ODEs are frequently phrased on . We collect recent results for neural ODEs on manifolds and present a unifying derivation of various results that serves as a tutorial to extend existing methods to differentiable manifolds. We also extend the results to the recent class of neural ODEs on Lie groups, highlighting a non-trivial extension of manifold neural ODEs that exploits the Lie group structure.
25 March 2026 in Communications in Nonlinear Science and Numerical Simulation
Neural ordinary differential equations (neural ODEs) are a well-established tool for optimizing the parameters of dynamical systems, with applications in image classification, optimal control, and physics-learning. Although dynamical systems of interest often evolve on Lie groups and more general differentiable manifolds, theoretical results on neural ODEs are frequently phrased on $\R^n$. We collect recent results on neural ODEs on manifolds, and present a unifying derivation of various results that serves as a tutorial to extend existing methods to differentiable manifolds. We also extend the results to the recent class of neural ODEs on Lie groups, highlighting a non-trivial extension of manifold neural ODEs that exploits the Lie group structure.