Modeling and Control of a Pneumatic Soft Robotic Catheter Using Neural Koopman Operators
Yiyao Yue, Noah Barnes, Lingyun Di, Olivia Young, Ryan Sochol, Jeremy DeLaine Brown, Axel Krieger
AI summary
Problem
Accurate modeling and control of pneumatic soft robotic catheters is hindered by complex nonlinear dynamics and limited sensing, making reliable open-loop navigation without continuous imaging feedback difficult.
Approach
The authors develop an end-to-end neural Koopman operator framework that automatically learns flexible state and input lifting representations from data, integrated with model predictive control for open-loop trajectory planning.
Key results
- Jointly learns flexible lifted-space representations and Koopman operator end-to-end
- Achieves average targeting errors of 2.1±0.4 mm in position and 4.9±0.6° in orientation
- Outperforms monomial basis Koopman, linear state-space, and PCC baselines in accuracy and efficiency
- Enables reliable open-loop control for a simulated cardiac ablation task without continuous imaging feedback
Why it matters
Provides a robust, data-driven control framework that enhances navigation precision and minimizes radiation exposure for minimally invasive cardiac interventions.
Abstract
Catheter-based interventions are widely used for the diagnosis and treatment of cardiac diseases. Recently, robotic catheters have attracted attention for their ability to improve precision and stability over conventional manual approaches. However, accurate modeling and control of soft robotic catheters remain challenging due to their complex, nonlinear behavior. The Koopman operator enables lifting the original system data into a linear “lifted space”, offering a data- driven framework for predictive control; however, manually chosen basis functions in the lifted space often oversimplify system behaviors and degrade control performance. To address this, we propose a neural network-enhanced Koopman operator framework that jointly learns the lifted space representation and Koopman operator in an end-to-end manner. Moreover, motivated by the need to minimize radiation exposure during X-ray fluoroscopy in cardiac ablation, we investigate open-loop control strategies using neural Koopman operators to reliably reach target poses without continuous imaging feedback. The proposed method is validated in two experimental scenarios: interactive position control and a simulated cardiac ablation task using an atrium-like cavity. Our approach achieves average errors of 2.1±0.4 mm in position and 4.9±0.6◦in orientation, outperforming not only model-based baselines but also other Koopman variants in targeting accuracy and efficiency. These results highlight the potential of the proposed framework for advancing soft robotic catheter systems and improving catheter- based interventions.