Research On Collision Detection Algorithm In Robot-assisted Laparoscopic Virtual Surgery

In recent years, robot-assisted virtual surgery continues to evolve and itsimportance is increasing in the medical field. Robot-assisted laparoscopic virtualsurgery can greatly help doctors design surgery projects and train surgical skills. Ina virtual surgery simulation system, collision detection is a prerequisite forcalculating feedback force and soft tissue deformation. It can improve the user’ssense of immersion and system fidelity. However, most algorithms are concentratedin collision detection between rigid objects. Research on collision detection forvirtual scene including soft tissue is relatively rare. It becomes one of the bottleneckproblems of virtual surgery simulation system. Collision detection problemcombines contents of computer graphics, biomechanics, material mechanics andmany other subjects, becoming a research hotspot. Therefore, carrying out theresearch on collision detection in robot-assisted laparoscopic virtual surgery hasimportant theoretical significance and application value.The virtual surgery scene has two features, namely complexity of the surgicalinstrument end movement and deformation of soft tissue. Collision detectionalgorithm based on changeable direction hull is proposed by analyzing thecharacteristics of classical hierarchical bounding box and also combining thefeatures of the virtual scene with fixed direction hull method. According to itsspatial geometry, building scheme for a single changeable direction hull boundingbox is designed. Construction project for bounding volume trees is analyzed fromdegree of the tree, construction methods and principles of nodes splitting. Collisiondetector is finally built for objects. In order to observe the establishment effect ofbounding box trees, visualization for construction process is realized.Global scheme for collision detection of the virtual scene is designed andintersection test function call for each node is implemented recursively. Necessaryand sufficient condition for two bounding box intersection is analyzed and thetesting method is designed. The test sequence is also elaborated. In addition,primitives included by intersecting nodes are tested, and the collision result isfinally obtained.Compared with the intersection test, the time of updating bounding box treestakes over a large proportion in collision detection and trees updating is the keyproblem for the research. According to the motion parameters of master hand andthe structure of surgical instruments, kinematics calculation of the instrument iscompleted. The method for updating bounding volume trees of tumbling objects is designed. The physical model of soft tissue is built and soft tissue deformation isimplemented combining with collision detection algorithm. Bounding volume treesupdating after deformation is mainly studied. The key issues of changeabledirection hull algorithm are analyzed and elaborated in detail.Finally, the simulation platform is built and the simulation experimentsincluding collision information visualization, interaction of soft tissue deformationand contrast between changeable direction hull with the classical algorithms aredesigned. Experimental results show that detailed collision information is obtained,realistic soft tissue deformation is presented, and the proposed algorithm hasachieved good improvement effect in the tightness of the bounding box tree and therapidity of collision detection.In this paper, a changeable direction hull collision detection algorithm isproposed. The theory has been applied for collision detection between the surgicalinstruments and the liver model, and the results show that it can realize the desiredeffect. It provides theoretical and application basis for force feedback and soft tissuedeformation in virtual surgery.