Background Traditional methods in cardiothoracic surgery require large incisions such as sternotomies. Recent work has demonstrated that robotic systems enable execution of coronary artery bypass graft (CABG) procedures through small incisions. In practice, these systems have proved cumbersome to use, with shortcomings that include decreased visual and haptic information, motion constraints, and the need for cognitive spatio-motor remapping from the surgeon’s hands to the instrument space. This necessitates extensive training, slows the execution of procedures, and adds to the surgeon’s workload. | ![[pic]](zeus_overall.jpg) |
![[pic]](surgeon_at_console.jpg)
|
Virtual Fixtures for MIS We propose to help alleviate these difficulties through virtual fixtures. A virtual fixture is a computer-generated constraint that simplifies task execution by reducing precision requirements or the number of degrees of freedom that must be controlled. On a desktop computer, these fixtures are analogous to computer mouse features such as snap-to-grid that make it simpler to precisely position the cursor. In this project, the validity of the concept of virtual fixtures is tested on the ZEUS surgical robot system (Computer Motion, Inc., Goleta, Calif.). In this project, we will develop virtual fixtures and surgical macros for the IMA harvest portion of robot-assisted CABG procedures. The patient will undergo a CT scan before surgery. Small metal pins will also be inserted percutaneously between the ribs, to provide fixed landmarks for reference during surgery. The resulting image set will be processed to define the location of the artery relative to the registration pins, and if possible, the location of each major branch of the artery will be determined. In surgery, the surgeon will bring the tip of a robot-mounted calibration instrument into contact with each of the pins. This will permit the robot to determine the location of the pins, and thus the location of the artery from the CT image data. In our initial implementation, a virtual fixture will constrain the instrument’s motions, as commanded by the surgeon, to appropriate paths adjacent to the artery. In the next stage, a surgical macro will move the robot along the path adjacent to each artery to dissect it free of the chest wall. The procedure to be developed will involve the following steps: |
![[pic]](animal_closeup3.jpg)
![[pic]](virtfix.gif) | Before surgery: 1. Insert registration pins between ribs near interior mammary artery (IMA) 2. CT Scan of the patient 3. Process image to determine location of IMA and registrations pins in CT data 4. Generate robot controller instructions: virtual fixture locations or surgical macro paths During surgery: 5. Establish patient location by touching instrument tip to each registration pin 6. Invoke virtual fixture or surgical macro and dissect IMA from chest wall |
· S. Park, R.D. Howe, and D. Torchiana, "Virtual Fixture for Robot-Assisted Minimally-Invasive Cardiac Surgery," to be presented at MICCAI 2001
· F. Lai and R.D. Howe, "Evaluating Control Modes for Constrained Robotic Surgery," IEEE International Conference on Robotics & Automation, San Francisco, April 2000.
· F. Lai, R.D. Howe, P.A. Millman, and S. Sur, "Frame Mapping And Dexterity For Task Performance In Robotic Endoscopic Surgery," in N. Olgac, ed., Proc. of the ASME Dynamic Systems and Control Division, ASME IMECE, Nashville, Nov. 14-19, 1999, DSC-Vol. 67.