A Study on the Capsule Robot Control Simulation System Based on Complex Magnetic Field Coupling

Capsule robots show great promise in biomedical engineering by facilitating internal diagnostics and therapeutics. Existing capsule robot control techniques, which usually depend on magnetic field uniformity, are susceptible to external environmental influences. This considerably restricts their adaptability in complex and dynamic settings. As a result, control strategies based on reinforcement learning have emerged as a significant area of study. Nevertheless, real-world capsule robot control systems face challenges in implementing reinforcement learning training. To address this, this study develops a capsule robot control simulation system using the Unity 3D platform. Integrating manipulator control, complex magnetic field modeling with four coils, and capsule robot motion simulation, the system can create intricate experimental scenarios with flexibility. Using the Comsol Multiphysics software for finite element (FEM) simulation of complex magnetic fields generated by coils, combined with Unity 3D real - time rendering technology, dynamic control and motion simulation of capsule robots in a virtual environment are achieved. Experimental results demonstrate that the system can effectively simulate complex magnetic field coupling with four coils and enable steering and movement of capsule robots through manipulator adjustments and coil current control, thus offering strong support for reinforcement learning training.