Precise motion control of concentric-tube robot based on visual servoing

Abstract

Composed of multi-section precurved tubes, continuous concentric-tube robot(CTR) has the potential of reaching surgical target during minimally invasive surgeries. Since concentric tubes are made of super-elastic nickel-titanium alloy, they can present different shapes when they extend and rotate with respect to each other. Compared to traditional surgical robots, CTR is superior in small size and flexible bending so that it can work in tiny space and adapt well to non-structural environment with multiple obstacles. In this paper, we proposed a CTR with 38cm in length and 1.85kg in weight which is assembled on a two-freedom platform to enable surgeons to perform various surgeries. Structure design of CTR and an image-guided algorithm are described in detail. For the propose of controlling robot to reach the target more precisely, a single camera is mounted at the tip of concentric tubes which can estimate the mapping between image plane and robotic joint space. This paper presents a model-less method based on Kalman filter to online estimate image Jacobian matrix, therefore robot is capable of capturing the target accurately accompanied by position changes on image plane. Experimental demonstration is also presented in this paper to verify the proposed method.