Modeling and Visual Servoing of Single and Cooperative Continuum Robots

This thesis dealt with modeling and visual servoing (VS) of continuum robots (CRs) and one of their recently introduced configurations, cooperative-CRs (CCRs). Application-wise, the focus was on small-sized CRs which are a commonly used type of CRs in minimally invasive surgeries (MIS).  First, an introduction to continuum robots, and various configurations of CRs were presented. In the next step, kinematic modeling of single-CR and CCRs were presented. Furthermore, dynamic modeling of smallsized CRs was studied. Tendon-driven small-sized CRs include continuous interaction between tendons and backbones, which introduce their own challenges into the modeling.    To control the system, firstly, a visual predictive control (VPC) system was employed for a single CR. Many scenarios were simulated to investigate the performance of the control scheme including positioning tasks in the presence and absence of various actuation and visibility constraints, a comparative study between VPC and image-based visual servoing (IBVS) approaches, the effectiveness of the parameter of prediction horizon (Np) on the system performance, the effectiveness of depth estimation and Np on the error and convergence time, and system robustness to the noise of actuators and imaging process. Finally, the system performance was experimentally verified.  

In the next step, a new switching VS control approach was developed to control CCRs. As the CCR model is based on Cosserat rod approach, the corresponding differential kinematics is obtained numerically which may result in ill-behaved or computationally expensive Jacobian. To address this issue, a scaled and nondimensional IBVS (SND IBVS) control scheme was introduced which is used to navigate the system at the beginning of the positioning task. The stability of the system is also proved in the sense of Lyapunov. To study the performance of the proposed switching controller, various scenarios were simulated.      

Moreover, an experimental set-up for the automation of CRs was developed to be used as the platform for verification of the system controller and to be served as a test-bed for the future studies of this research.