Abstract
This thesis proposes a robotic assistant for MRI-guided percutaneous liver interventions. The main objective for this assistant is to facilitate the surgical procedures and fulfill the constraints of compatibility, lightness and compactness imposed by the MRI environment. An additional requirement for this assistant is its adaptation with the physiological motion of organs produced during the breathing cycle. The limitation of the state-of-the-art robotic assistants to cope with the additional requirement motivated us to propose in this thesis the use of tensegrity mechanisms and their inherent capacity of reconfiguration and stiffness variation in order to overcome the various design challenges.Our contributions include the design, the modeling and the control of tensegrity mechanisms. Two approaches to control the position and the stiffness of the robot are introduced. The first one is based on the implementation of a tension distribution algorithm. The second one is based on a linear model predictive controller. The position tracking and the stiffness modulation performances are evaluated on two versions of the prototype using the proposed control approaches.