Study on Continuum Robots for Minimally Invasive Surgery


Student thesis: Doctoral Thesis

View graph of relations


Related Research Unit(s)


Awarding Institution
Award date14 Jan 2021


Minimally invasive surgery (MIS) that involves the utilization of surgical instruments through small incisions has been increasingly applied as an alternative to open surgery due to its improvement in safety, recovery time, and postoperative complications. Continuum robots that enjoy the feature of slenderness and mechanical compliance can circuitously reach hard-to-reach target sites to reduce risks during operation. Therefore, they are highly suitable for MIS and have been increasingly adopted in hospitals. Although considerable progress has been achieved, continuum robots still suffer from difficulties in modeling, sensing, and control, particularly when they work in constrained environments because interaction is inevitable between continuum robots and their environment. In this study, two types of continuum robotic systems were developed and investigated for automatically performing MIS tasks in constrained environments. This thesis includes the following two parts:

In the first part, a continuum robotic system with a hyper-redundant structure was developed to deform a soft object in a constrained environment. A corresponding actuation mechanism was also designed to deploy the tendons. The kinematic model of the system was analyzed on the basis of the hyper-redundant structure. To control the continuum robot to deform a soft object in a constrained environment, a novel vision based model-free control strategy was proposed. In contrast with most deformation control approaches, our proposed method can deform a soft object into a desired shape without the need of identifying the object’s deformation model. The proposed method included an online dynamic estimator for approximating the deformation Jacobian matrix that associates the actuator input with the deformed output of a soft object in real time, avoiding any model identification steps. A model predictive controller with a reference trajectory was further developed to ensure smooth operation. Simulations and experiments were performed to verify the effectiveness of the proposed approach in deforming a soft object in a constrained environment.

In the second part, a 3-mm-diameter continuum robot that can be inserted into an endoscope was designed to perform laser beam steering. A backlash rejection actuation mechanism was designed to address the problem of backlash when actuating the bending section of the robot in push-pull mode. A piecewise constant curvature model was applied to analyze the kinematics of the robot. In addition, a robust control approach that enables to handle noise from visual feedback and disturbances from the environment was proposed to control the flexible continuum robot for laser beam steering. The proposed control method for the first time considered the deterioration of visual feedback from an endoscope. To overcome noise from visual feedback and disturbances from the environment during operation, a local Jacobian matrix that maps the actuation space to the image space was approximated online by using the sensing data stored in the limited memory, where the influence of outliers can be filtered. A linear state observer was also designed to compensate for estimation errors. Then, a second-order sliding mode controller was developed to achieve robust control of the flexible manipulator during laser beam steering.

In summary, this thesis reported the design and control of two continuum robotic systems: one for automatically deforming a soft object and another for steering a laser beam. The development of the continuum robotics systems and the corresponding control methods exhibited the advantages of precise manipulation in unknown constrained environments. This study will bring continuum robots to automatically assist surgeons and/or perform a single surgical step or several connected steps of surgical procedures during MIS.