A Fast Collision Detection Algorithm Between Multibodies

Collision detection, also refers to as contact detection, is a computational method to judge whether two or more objects intersect with each other. It is being wildly used in the field of industry, video game and etc. While it is being extensively applied as a computational method, it is more related to computational geometry and is studied by the researchers of computer graphics, failing to combine a more accurate physical and mechanical simulation solver to analyze the objects’behavior after the collision happened. Meanwhile, in many existing commercial mechanical software, the information of the contact surface has to be input in advance by the mechanical backgrounded users to get a collision or contact analysis. In such case, in dealing with multibody contact issue, all the surface has to be set as contact surface and the software will bear heavy computational cost to find the contact pair and the contact points. Thus, this thesis introduced a program of collision detection adapted to multibody issues, enabling to find the contact points before FEM analysis.In this thesis, the input information to solve a FEM problem is studied, the one that is required by collision detection is been retrieved to use as the input interface. Then output information needed by the software Ferview to get the image displayed is also studied, thus the program’s output interface is set. Then the structural of Prof. Feng’s self-developed FEM software is analyzed and a way to embed the collision contact algorithm has been found. This thesis has also summarized the existing mainstream collision algorithms, and discussed two of them in detail:one by spatial partitioning, the other by bounding volume hierarchies (Spheres, Axis-aligned bounding boxes) and the program based on these two kinds of algorithms have been developed. The structural construction of the program is presented and some important parts of the code is explained in detail. A few examples showed the result of the collision detection program and the results when the collision detection program is combined to a contact-issue-solving program. The functionality of the program is verified. And some comparisons between these algorithms is made.In the last part of the thesis, some forecast of the algorithm is made, and some thoughts and possible ways about how to optimize the algorithm is proposed.