Semiautonomous Robotic Manipulator for Minimally Invasive Aortic Valve Replacement

Figures

• 

  Fig. 1. Schematic of MIAVR by the proposed robot. (a) Illustration of the procedure operated by a surgeon with the help of an endoscopic view and haptic interface to reach the intervention site (phase I). (b) Exploring and selecting the commissure points (i.e., the space between each anchored leaflet and the aortic wall) in the endoscopic views by the surgeon (phase II). (c) Automatic centering of the tip and introducer translation to the expansion site based on image data (phase III).

• 

  Fig. 2. System overview. (a) Geometrical considerations for robotic minithoracotomy MIAVR. (b) Robot overview and its structures showing cable actuation and sensing principle, introducer, and flap actuation. (c) Manipulator in a bent configuration and its kinematic modeling principle. (d) System mechanics and external load modeling principle. (e) Robot motion planning to access the aortic valve.

• 

  Fig. 3. Introducer mechanism for converting the spring rotation to linear translation of the valve. This valve is crimped and held inside the manipulator by a custom cartridge, and it is charged in the robotic introducer before entering the body.

• 

  Fig. 4. Surgical navigator developed thanks to the integrated tip cameras. (a) Calibration tool with the reference points marked with different colors to facilitate their identification in the camera views. (b) Simplified problem associated with the triangulation of the 3-D U point, based on the knowledge of its projection on the camera image planes, ui and ui+1.

• 

  Fig. 5. Control and automation system architecture. Blue boxes represent the software part of the system, and green boxes are the hardware part.

• 

  Fig. 6. (a) Comparison of the final manipulator shape (solid line) and its tip motion path (dashed line) for different horizontal distances xBE of the manipulator base from the insertion point E. The motion stops when the manipulator tip aligns with the target position T. (b) Required introducer length lI to reach the target point and the mean and maximum follow-the-leader error values versus xBE. (c) Larger the distance xBE, the lower the follow-the-leader error but larger required introducer length. The manipulator z-axis positive direction is upward.

• 

  Fig. 7. Experiments with simplified prototypes to validate the system. (a) FK and IK validation setup (inset: weights added to the introducer body). (b) Setup for validating the introducer translational model while the tube mock-up can rotate and provide various coil attachment lengths.

• 

  Fig. 8. Experiments to validate the system. (a) FK, IK, and FSO validation setup; inset: force observer validation by four weights hanging from the manipulator tip. (b) Setup for validating forces on the body of the manipulator during insertion. (c) Setup for measuring the accuracy of the 3-D center estimation while the target board was repositioned by the red arrows. (d) Setup for validating the controller’s accuracy in reaching a target point. (e) Experimental setup for performing artificial valve delivery in a silicony mock-up. (f) Points in Sorin Perceval and commissure points in the aortic root that should match for an ideal positioning.

• 

  Fig. 9. Results of the FM framework for the robot workspace (experimental cases 1).

• 

  Fig. 10. Results of the IM framework for experimental case 2. The accumulated error due to CC kinematics results in a large error in predicting the introducer linear motion.

• 

  Fig. 11. Results of the hybrid FSO framework for experimental case 2 with robot motion when external forces were applied. The force estimation results are relatively accurate in the z-axis direction. The accumulated error due to CC kinematics and the high stiffness of the system result in predicting unrealistically large lateral forces in the x − and y-axis directions.

• 

  Fig. 12. Results of the IM framework for experimental case 3.