Optimal path, orientation and trajectory planning along arbitrarily shaped surfaces for image-based robot-automated medical procedures

Abstract

Robot-assisted surgery offers many advantages in terms of precision and facilitation in medicine, whereby the physician controls the system online. However, operational errors cannot be ruled out and the traversal along complex surfaces, e.g., to perform precise milling or cutting, is still extremely complicated. This contribution introduces a physician-planned and robot-executed image-based automation for uniform motion along arbitrarily shaped surfaces, which will especially improve the accuracy in surface-true medical procedures, e.g., executing precise cuts or removing adherent tissue in the case of spinal stenosis. The physician marks desired waypoints on relevant objects that are previously segmented and meshed from the computer tomography (CT) or magnetic resonance imaging (MRI) scans. A length-optimal path planner determines the shortest path along the surface in 3D space by a two-stage optimization, supplemented by a surface adapted orientation strategy for correct alignment of the end-effector. Marking significant points rather than entire paths on curved surfaces simplifies the applicability in 3D imaging for the physician. Subsequently, a node-faithful trajectory planning algorithm for uniform motion based on the path is introduced. Using robot specific inverse kinematics, the subsequent transformation into the joint space takes place. Thus, the approach can be transferred to common systems existing in medical robot assistance. The execution of the monitorable surgical planning procedure is performed both in simulation and experiment on a complex shaped 3D-printed lumbar vertebra from a CT scan using a Stäubli TX2-60 manipulator.