Active Disturbance Rejection Control With Data-Driven Compensation for a Novel Extensible Continuum Robot

摘要

Continuum robots (CRs) are widely used in minimally invasive surgery due to their inherent flexibility, compliance, and ease of miniaturization. To overcome the limitations of existing CRs in terms of operation dexterity, we design a novel extensible CR that combines spring structures and rod actuation. This design features a slim profile, large working channels, and the ability to independently control both bending length and curvature, enabling a wide range of configurations and enhanced dexterity. On the other hand, due to the nonlinear deformation and infinite degrees of freedom of the CR, controlling the robot presents many challenges, including nonlinearity and uncertainties. Thus, we propose an active disturbance rejection control with data-driven compensation (ADRC-DDC). This method uses a constant curvature model to capture the main dynamic behavior of the CR and combines data-driven disturbance estimation and disturbance estimation from an extended state observer (ESO). The proposed method can better handle system uncertainties, alleviate the burden on the ESO, and improve both robustness and control performance. Experimental results demonstrate the superiority of the designed extensible CR and the proposed ADRC-DDC method.