Wireless-controlled microrobots/micro-particles can be manipulated in living bodies to solve numerous medical problems, such as drug or cell delivery, minimally invasive surgery, and blood clot removal. Recently, the use of magnetic fields to manipulate microrobots has received extensive attention because of its precise controllability, strong driving ability, and sound operating safety in the in vivo environment. The magnetic force generated by the electromagnet can be easily adjusted by changing the current input of the electromagnetic coil, thereby providing greater flexibility and safety in use. In biomedical applications, the precision of microrobots’ motion determines the therapeutic effect directly. Therefore, the manipulation of microrobots should meet the accuracy requirements. Automatic control based on visual feedback can realize high-accuracy manipulation. This thesis focuses on how to control a magnet-driven microrobot/microrobots swarm automatically.

First, a four-coil electromagnetic manipulation system is used to control the movements of microrobots in a 2D space. After dynamic modeling, the moving trajectory of the microrobot is designed on the basis of an artificial potential field. An estimator for the position is then developed with stability analysis by a Lyapunov approach. A super-twisting algorithm is further applied to control the microrobot to move along the desired trajectory. The proposed method can guide the microrobot to move automatically in a simulated vascular structure.

Second, a six-coil electromagnetic system is designed to realize automated mapping and path planning in a 3D workspace. A path planner is designed to search for an optimal path in a 3D space with obstacles automatically, and a cascaded control algorithm is developed to control the microrobot’s movement along the planned path. The path is generated by combining the A-star and minimum jerk methods. The collision of the generated path with the obstacles is prevented, and the hysteresis caused by the current change is minimized by reducing the jerk of the movement with a passable path from the starting point to the endpoint. In the cascaded control algorithm, the incremental nonlinear dynamic inversion method and proportional control are used as the inner and outer loops to guide the microrobot’s movement along the desired trajectory with an appropriate velocity while eliminating the influence of the system uncertainty and external disturbances.

Third, a novel electromagnetic system is designed to realize the control of the micro-particle swarm. Providing a sine-wave current input to one electromagnetic coil can gather the particles to form a vortex. The degree of concentration can be changed by adjusting the magnitude and frequency of the input current. Moving the coils in the space while maintaining a high concentration degree of the swarm can realize the control of the swarm. Due to the rotational motion, the control gain of the entire system is difficult to obtain, and some dynamics are difficult to model, that is, the system has parameter and model uncertainties. In this thesis, the sliding mode control is adopted, the error is converged to a certain range through the sliding mode surface, the robustness of the system is guaranteed, and the chattering is eliminated to a certain extent. Finally, the control input is restricted within a specific range to ensure the stable existence of the vortex.

In summary, the proposed electromagnetic manipulation systems provide valuable platforms for the study of microparticles’ manipulation. Moreover, the proposed control method can guide the microrobots to their desired position automatically. Simulations and experiments are performed to demonstrate the effectiveness of the proposed approaches. More importantly, the proposed strategy is aimed at in vivo applications, it can also be applied to many real-life scenarios, such as moving microrobots in body or tissue fluids, to treat of liver, stomach, and joint cavity diseases.