Dynamic Modeling an‎d Control of a Soft Intelligent Ferromagnetic Robot in Complex Environment

10/6/2026 12:00:00 AM

Recent advances in medical robotics have enabled the development of intelligent ferromagnetic robots that can be guided by external magnetic fields. Due to their high flexibility an‎d small scale, these robots can access complex an‎d narrow anatomical pathways, such as blood vessels an‎d other intricate structures. However, predicting an‎d controlling their deformation under magnetic fields remains a significant challenge due to their continuum nature an‎d high degrees of freedom. Achieving accurate an‎d safe operation in such applications requires the development of precise mathematical models an‎d advanced control strategies for the analysis an‎d guidance of these robots. In this dissertation, the modeling of intelligent ferromagnetic robots under the influence of external magnetic fields generated by electromagnetic systems is presented. Using Cosserat theory, the kinematic equations describing the robots’ deformation were derived, an‎d the equilibrium between position, orientation, internal forces, an‎d moments was solved. To validate the model, the simulation results of the magnetic robots were compared with COMSOL simulations an‎d recent experimental data. The comparisons showed that the proposed model predicts the deformations with high accuracy, with a prediction error along the robot length of less than 8%. Furthermore, the effects of key geometric parameters, including length an‎d cross-sectional area, on the deformation were investigated, revealing that variations in these parameters play a significant role in optimizing the performance of ferromagnetic robots. This quantitative accuracy an‎d analysis are particularly important for medical applications an‎d the safe guidance of these robots. Subsequently, the dynamic modeling of intelligent ferromagnetic robots under the influence of external magnetic fields is presented. The modeling is based on Cosserat theory an‎d geometric discretization of the robots, enabling precise analysis of their dynamic behavior an‎d more efficient implementation of numerical simulations. To validate the model, the numerical simulation results were compared with experimental data an‎d the findings from the first part of the dissertation. The comparison showed that the proposed model predicts the robots’ dynamic behavior with a prediction error of less than 7%, demonstrating good agreement with experimental data. These results indicate that the developed model provides a significant capability for analyzing an‎d designing the control of intelligent robots in complex medical applications. In the final part of the dissertation, a model-based controller for intelligent ferromagnetic robots is presented. This controller combines information from the kinematic an‎d dynamic models to predict the robot’s deformation an‎d control it in real time, enabling the execution of delicate medical tasks within complex anatomical environments. The controller design was carried out in two stages: first, a neural network was trained to map target positions to the corresponding deformations based on the kinematic model; then, a nonlinear sliding mode controller was developed to ensure stability an‎d accuracy in positioning the robot tip an‎d following the desired trajectories. Point-to-point an‎d trajectory-based simulation results demonstrated that the controller can track target positions with a prediction error of less than 2% relative to the robot length. This performance highlights the effectiveness of the proposed approach for precise, real-time medical applications.

ربات هوشمند مغناطيسي , مدل‌سازي سينماتيك , مدل‌سازي ديناميك , تئوري كاسرات , شبكه عصبي , كنترل غيرخطي

Ferromagnetic continuum robot , Kinematic Modeling , Dynamic Modeling , Cosserat Theory , Neural Networks , Nonlinear Control

Pouya Mallahi Kolahi

Moharam Habibnejad Korayem