Dynamics-Based Motion Control for a Hybrid-Driven Continuum Robot With Continuously Variable Stiffness

Abstract

While the hybrid driving method effectively addresses the contradiction between inherent compliance and the finite load-bearing capability of continuum robots, integrating multiple actuations poses challenges in modeling and control. This article introduces a dynamics-based robust uncertainty estimation and control (DRUEC) method for hybrid-driven continuum robots with continuously variable stiffness to tackle the fast internal dynamic variations and enhance motion tracking accuracy. Initially, a conventional kinematic formula is established to transfer all local force and position vectors into the global coordinate. Subsequently, the Euler-Lagrange methodology is employed to construct the entire dynamic model within the actuation space. For improving programming efficiency and independence from system parameter identification technologies, explicit expressions of matrices in the constructed dynamic model are derived by using the chain rule and properties of homogeneous coordinate transformation. In addition, a novel robust uncertainty estimator (RUE) is proposed to estimate modeling errors arising from the unmodeled dynamics and parameter perturbations. Various experiments are implemented based on a hybrid-driven continuum robot with two segments. Comparative results show the effectiveness of the proposed scheme over the classical methods. Note to Practitioners—The motivation for this article is to enhance the motion tracking accuracy of hybrid-driven continuum robots. Due to their inherent compliance, continuum robots exhibit a great adaptation capability to restricted environments. However, a conflict between intrinsic compliance and positioning accuracy of the endpoint limits their practical applications. To tackle this issue, the hybrid driving method offers an accessible solution by decoupling stiffness regulation from position adjustment. This enables different stiffness levels at the same posture, allowing the continuum robot to withstand varying loads without experiencing significant deformations. Nevertheless, the incorporation of multiple actuations also complicates modeling and control due to the increased coupling and nonlinearity. Moreover, the effective operation of continuum robots needs to timely deal with the effects induced by the rapid internal dynamic changes and relatively large movement speeds. These facts imply that the commonly used quasi-static models or kinematics-based methods are insufficient. Therefore, this article is dedicated to improving the control accuracy from the perspective of dynamics-based control approaches.