Real-time trajectory tracking for externally loaded concentric-tube robots

Abstract

Concentric-tube robots can offer a suitable compromise between force and curvature control. In a previous study by the authors, a real-time trajectory tracking scheme for an unloaded concentric-tube robot was developed. One of the practical barriers to the use of a concentric-tube robot in medical applications is compensation for the impact of environmental forces which can cause drastic deterioration in tracking performance. In this paper, by modifying the robot's forward kinematics and Jacobian, a new method is developed to facilitate tip tracking in real-time while accounting for an external load at the robot's tip. By considering the tip deflection resulting from the external load, a novel dual-layer control architecture is proposed to compensate for this deflection during trajectory tracking. In order to measure the force exerted on the tip position of the robot, a new technique is proposed that can move the sensing system from the distal tip to the proximal base. Experimental results are given to illustrate the effectiveness of the proposed method.

Autonomous Steering of Concentric Tube Robots via Nonlinear Model Predictive Control

“…Simulations compare MPC performance with Jacobian-based controllers, which to the best of authors knowledge, are the most common type of controller for CTRs [8], [9], [21]. The Jacobian can be used with the following closed-loop control law to steer the CTR [22]:…”

“…Rucker et al [8] proposed a numerical approach to increase the computational efficiency of Jacobian estimation for a CTR under external loading. Xu et al [9] proposed realtime estimation of the robot Jacobian by discretizing the robot into small sublinks. The Jacobian was used for closed-loop control of the CTR's tip.…”

“…Most recent research in CTRs has focused on their open-loop steering [1], [3]. State-of-the-art closed-loop control strategies rely on linear approximation of robot inverse kinematics near equilibriums [9], [12] but neglect the system's non-linearities, potentially leading to instability and reduced controller accuracy. Recently, researchers proposed application of dynamic active constraints that could improving the safety and stability also of CTRs [13].…”

“…In this paper, we design a novel controller for steering of CTR's end-effector. Unlike current CTR controllers [9], [11] that employ the local robot Jacobian, the Model Predictive Controller (MPC) uses future predictions of robot behaviour to handle explicit constraints on the robot state and control inputs, therefore being able to steer the robot away from unstable configurations commonly known as "snapping" points. Snapping is a mechanical instability in CTRs caused by the release of the elastic potential energy that is accumulated due to tube bending and twisting [1].…”

Autonomous Steering of Concentric Tube Robots for Enhanced Force/Velocity Manipulability

“…Most developed controllers assume that the CTR is stable by design and open-loop control strategies are used to steer the CTR [3], [12]. Closed-loop control strategies rely on the linear approximation of robot inverse kinematics near equilibrium [13]- [15] and try to avoid unstable paths via offline path planning. Further, limited effort has been invested on control that takes into account force application capabilities.…”