Scalable parallel collision detection simulation
SIP '07 Proceedings of the Ninth IASTED International Conference on Signal and Image Processing
Parallelizing a real-time physics engine using transactional memory
Euro-Par'11 Proceedings of the 17th international conference on Parallel processing - Volume Part II
Hi-index | 0.00 |
The automatic recognition of geometric constraints invirtual assembly and maintenance operations relies in thedetermination of intersecting surfaces between virtualprototypes. This is a key challenge in many virtualprototype applications, where it is necessary to findcollisions precisely and interactively. This paper presentsa novel algorithm to determine intersecting surfaces atinteractive speed in a virtual prototyping environment.The proposed algorithm is based on the Overlapping Axis-AlignedBounding Box (OAABB). The OAABB concept isused effectively to eliminate the number of surfaces thatcannot intersect and improve performance. The algorithmalso facilitates the implementation using parallelcomputing methods. OpenMP is used, taking advantage ofshared memory multiple processors and reducing theoverall time complexity of the collision detectionalgorithm. To achieve an efficient parallel simulation, it isnecessary to provide an efficient load balancing scheme.Our experiences in parallelising the code to achieve abetter work distribution are also described. Results showthat the proposed collision detection achieves interactiverates in real industrial applications as desired.