Collision Detection In Virtual Surgery Simulation System

With the rapid advancement of computer hardware and software technologies, virtual reality and haptics based systems are founding applications in various fields. For such a system, collision detection is the key technology to determine whether the virtual object has been collided in virtual scene. Precise and rapid collision detection algorithms play a vital role on the authenticity and interaction of the virtual reality. Thus, collision detection is now one of the hottest topics of current research.In thesis, we deal with the collision detection problem and develop a new algorithm, which is based on the classical AABB bounding boxes collision detection algorithm. We aim at surgical simulation. To improve the efficiency of the AABB collision detection algorithm and enhance the degree of realism of the virtual surgery simulation system and its real-time interaction with the users, the conventional AABB algorithm has been improved in several different aspects. Since the refreshment update of the simulated object (such as the virtual patient) is slow, a new optimization algorithm is used in the pretreatment process to accelerate the AABB tree's updating after object deformation; With the introduction of the temporal coherence concept, an layer-by-layer sorting algorithm is applied on the AABB tree to test the intersection among boxes. Thus, the computation complexity of the collision detection algorithm is reduced; The storage space of the AABB is compressed in the optimization process in order to save memory and reduce the space complexity; The concept of the instrument's effective active region has been introduced in order to reduce the scale of the AABB tree. An optimization algorithm, which employs a primitive list instead of the traditional binary tree, is developed to make the model of the instruments. As a result, the efficiency of the collision detection is improved significantly.Compared with the conventional AABB algorithm, the presented collision detection algorithm provides significant improvements in terms of both the space complexity and time complexity. The developed collision detection algorithm has been applied to a specific virtual surgery simulation system, and the simulation results show that the algorithm effectively improved the degree of authenticity and the realism of the interaction with virtual environments. Experimental results show the effectiveness and validity of the improved AABB collision detection algorithm.