Model-Free Adaptive Predictive Control for Trajectory Tracking of Magnetically Controlled Capsules

In this study, the model-free adaptive predictive control approach is incorporated into the magnetically controlled capsule inspection regime. Given the highly nonlinear and considerably time-delayed nature of the magnetically controlled capsule system, the compact form dynamic linearization technique is employed to convert the system model into a dynamic linear data model. Subsequently, grounded on this data model of the system, a model-free adaptive predictive tracking control algorithm is devised. This proposed control methodology dispenses with the need for precise system model details, relying solely on the force particulars at the robotic arm’s extremity and the positional feedback from the magnetic positioning apparatus, representing a data-driven control strategy. Experiments were conducted on a simulated magnetically controlled capsule testing platform, and the outcomes validated the efficacy of this method.