Research Analyzer
← Back ICRA 2026

Model Predictive Control with Reference Learning for Soft Robotic Intracranial Pressure Waveform Modulation

Fabian Flürenbrock, Yanick Büchel, Johannes Köhler, Marianne Schmid Daners, Melanie N. Zeilinger

PDF

AI summary

Key figure (auto-extracted from paper)
A learning-based MPC framework with Bayesian optimization rapidly learns optimal motor trajectories to safely and accurately modulate intracranial pressure waveforms.
Model predictive control Bayesian optimization Intracranial pressure Soft robotics Reference learning Neurological disorders

Problem

Existing soft robotic systems for intracranial pressure modulation rely on manual PID tuning and lack safety constraints, making accurate waveform control time-consuming and risky.

Approach

The authors combine an offset-free model predictive controller with a disturbance observer for precise motor tracking and a Bayesian optimization algorithm that automatically learns the optimal motor reference trajectory in real-time.

Key results

  • 83% and 73% reduction in mean and maximum tracking errors versus PID
  • Bayesian optimization converges to optimal motor trajectories in under 20 iterations
  • Successful in vitro replication of target intracranial pressure waveforms
  • Real-time offset-free tracking with explicit safety constraint handling

Why it matters

Provides a safe, automated control strategy for advancing research and therapies targeting cerebrospinal fluid dynamics and neurological disorders.

Abstract

This paper introduces a learning-based control framework for a soft robotic actuator system designed to modulate intracranial pressure (ICP) waveforms, which is essential for studying cerebrospinal fluid dynamics and patho- logical processes underlying neurological disorders. A two-layer framework is proposed to safely achieve a desired ICP wave- form modulation. First, a model predictive controller (MPC) with a disturbance observer is used for offset-free tracking of the system’s motor position reference trajectory under safety constraints. Second, to address the unknown nonlinear dependence of the ICP on the motor position, we employ a Bayesian optimization (BO) algorithm used for online learning of a motor position reference trajectory that yields the desired ICP modulation. The framework is experimentally validated using a test bench with a brain phantom that replicates realistic ICP dynamics in vitro. Compared to a previously employed proportional-integral-derivative controller, the MPC reduces mean and maximum motor position reference tracking errors by 83 % and 73 %, respectively. In less than 20 iterations, the BO algorithm learns a motor position reference trajectory that yields an ICP waveform with the desired mean and amplitude.

Index terms

Medical Robots and Systems Machine Learning for Robot Control Optimization and Optimal Control

Related papers