Learning Rate of the Model Algorithm for Iterative Learning Control in Lung Nodule Surgical Continuum Robot Systems

Abstract

During surgical operations, the distal end of lung nodule surgical robots is frequently confronted with diverse and intricate disturbances, thereby posing significant challenges for nonlinear control of such continuum robot systems. The continuum robot has a complex nonlinear dynamic model, and the coupling between the joints will affect each other, which makes the joint control of the continuum robot difficult. In addition, the motion of the continuum robot also needs a real-time control strategy. Based on the above analysis, this paper proposes a nonlinear iterative learning method, which is grounded in model algorithmic learning rates, for the control of the distal end of a surgical robot utilized in pulmonary nodule operations. This method not only considers the control error and its higher derivative, but also includes the parameters of the system model. Then, based on the learning rate determined by the model algorithm and the actual control input from the current iteration, the control input for the next iteration is calculated, thereby advancing the iterative learning process. Finally, the stability of the entire nonlinear iterative learning process is proved by the spectral radius condition under the global Lipschitz condition. The effectiveness and robustness of the proposed method have been verified through MATLAB/Simulink, demonstrating high precision and superior performance.