Continuum Robots: A Real-Time Model-Based Data-Driven Nonlinear Controller

Model-based control for continuum robots is often based on the piecewise constant curvature (PCC) assumption. However, the actual physical motions of continuum robots deviate significantly from this assumption, posing challenges to achieving the desired stability and accuracy in control. In this article, a real-time model-based data-driven controller is designed to enhance the precision and stability of continuum robots. An algorithm is designed to construct the static model of the cable-driven continuum robots. A real-time data-based kinematics model is formulated based on the static model, and the multistate motions of the robot are described in this model. A nonlinear control strategy is proposed based on the data-based kinematics model. Validation through simulations and experiments demonstrates the method's robust stability and impressive precision, achieving a positioning accuracy of 0.0318 mm and a tracking accuracy of 0.0434 mm. This article provides a general paradigm for model-based control of continuum robots, which is crucial to the advancement of this industry.