Modeling and analysis of a modular-structured tendon-driven continuum robot for the three-dimensional end effector coordinate

Abstract

This article introduces the design and control of a tendon-driven continuum robot (CR) segment with a modular structure consisting of a series of backbone discs connected in series. Thanks to this developed design; The backbone length can be changed by increasing or decreasing the number of discs, the number of segments can be increased thanks to the hollow flexible central backbone, and additional apparatus such as a camera or holder can be sent to the end effector. Multi-part and multi-degree-of-freedom CRs are difficult to control because their kinematics and dynamics are non-linear. To overcome this challenge, the development of a finite element method (FEM) based model is presented. In this framework, a nonlinear FEM formulation was applied to model large displacements and contact problems in CRs with motion control by a tendon driven mechanism. The study was supported by a series of experiments to evaluate the accuracy and performance of the developed model. A Feed-forward Artificial Neural Network (FNN) model was developed using the data obtained from the FEM model. The FNN model predicts the X, Y, Z coordinates of the end effector in the 3D plane, based on the tendon drive distance of the CR. The results demonstrated the accuracy and reliability of the proposed nonlinear FEM formulation. At the same time, the FNN algorithm developed with optimum parameters demonstrated that the X, Y and Z position of the end effector has a high estimation accuracy of R=0.9995 on average, depending on the amount of tendon drive.