Nonlinear dynamic modeling and model-based AI-driven control of a magnetoactive soft continuum robot in a fluidic...

Area of Science:

• Robotics
• Biomedical Engineering
• Control Systems

Background:

• Magnetoactive soft continuum robots (MSCRs) offer promise for biomedical applications due to their multimodal locomotion.
• Accurate modeling and control are crucial for steering MSCRs within complex biological environments.
• Understanding the coupled magneto-mechanical behavior and fluidic interactions is essential for MSCR development.

Purpose of the Study:

• To develop a novel nonlinear dynamic model for MSCRs that accounts for fluidic damping and drag forces.
• To create an AI-driven, model-based control method for precise MSCR navigation in simulated biofluidic environments.
• To validate the model and control strategy through experimental assessments and comparisons.

Main Methods:

• A nonlinear magneto-viscoelastic dynamic model was developed, incorporating damping and drag forces from fluid flow.
• A fractional-order sliding mode control (FOSMC) algorithm, integrated with deep reinforcement learning (DRL), was designed for trajectory tracking.
• A hardware-in-the-loop experimental framework was established to test the DRL-FOSMC algorithm under various conditions.

Main Results:

• The developed dynamic model showed accurate correlations between theoretical predictions and experimental data for MSCR responses.
• The DRL-FOSMC algorithm demonstrated enhanced tracking performance and reduced chattering in simulated fluidic environments.
• Experimental case studies confirmed the superiority of the proposed DRL-FOSMC algorithm over existing control methods.

Conclusions:

• The novel dynamic model accurately captures the behavior of MSCRs in fluidic environments.
• The AI-driven DRL-FOSMC provides effective and precise control for navigating MSCRs in simulated biological conditions.
• This research advances the potential of MSCRs for future biomedical applications requiring accurate in-body steering.