Research On Collision Detection Technology For 3D Simulation Of Industrial Robot

Collision detection is the key issue in the field of robot simulation, virtual reality and computer visualization, its basic task is to examine the connection and penetration between two or more objects. With the development of robot simulation technology, higher requirement is needed in simulation environment. Real-time and accurate collision detection plays an essential role in enhancing the facticity and immersion of simulation environment. This paper studies collision detection between robot and surroundings including robot itself in simulation process of industrial robot so that achieving the precise collision detection in three-dimensional simulation platform.Firstly, RAPID algorithm is improved to propose a hybrid hierarchical bounding box collision detection algorithm based on AABB and OBB. This paper constructs each object's AABB for its simpleness and fast intersection test to rapidly exclude objects that do not intersect in the initial stage of the collision detection. Furthermore, we construct each part's OBB for its good tightness and accurate intersection test to do the accurate collision detection. Construction method from top to bottom is adopted to construct AABB-OBB hybrid bounding volume hierarchy trees. After the collision detection experiment of two teapot models, the collision detection algorithm is found to have higher accuracy and stability, but the speed and storage can be further optimized.Secondly, this paper optimizes speed and storage requirements of AABB and OBB hybrid hierarchical bounding box collision detection algorithm. Speed optimization:the algorithm based on global search algorithm of axis divided theory is proposed to do the intersection test of AABB, and accelerated traversal algorithm based on spatio-temporal correlation is adopted to accelerate the travel of hybrid hierarchical bounding volume tree. Storage optimization:the method based on compressed storage root nodes is adopted to optimize storage of root nodes and based on fast triangle intersection test algorithm is adopted to optimize storage of leaf nodes. Experimental results show that optimized collision detection algorithm has higher efficiency and lower storage consumption.Finally, this paper constructs work scene of six degrees of freedom industrial robot named BUAA and positioner. In the process of the robot trajectory planning, it achieves precise and fast collision detection between robot and surroundings including positioner and ground and among robot joints.