N. Garica-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, May 1994.  Home/Search   Document Not in Database   Summary   Related Articles   Check

....necessary. Thus, the key issue for repositioning of bone fragments is collision detection, which is a fundamental problem in 3D interactive applications. In this context, most of the previous works focus on the interaction of polygonal objects, such as polygon meshes, polyhedra or splines (e.g. [13, 3, 6]) Unfortunately, those approaches suffer from the fact, that explicit surfaces are required, which can lead to unprecise results. A more natural approach for collision detection would be a solution for volumetric objects, since the containing voxels can describe different interaction behaviors of ....

N. Garica-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, May 1994.

.... that reconstruct new images from the reprojected pixels of reference images [22] Bounding boxes are also common in algorithms for modeling, from techniques that define complex shapes as Boolean combinations of simpler shapes [17] to techniques that verify the clearance of parts in an assembly [10]. Animation algorithms also exploit bounding boxes, especially collision detection algorithms for path planning [20] and the simulation of physically based motion [2, 19, 24] While empirical evidence demonstrates that the bounding box heuristic improves performance in practice, the goal of ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36--43, 1995.

....of the representation is only necessary in the parts where collision may occur. Space and object partitioning representations for collision detection are surveyed next. Octrees [37] or octree like structures [4] BSP trees [62] brep indices [9] tetrahedral meshes [49] and regular grids [31] are all examples of spatial partitioning representations. By dividing the space occupied by the objects, one needs to check for contact between only those pairs of objects (or parts of objects) that are in the same or nearby cells of the decomposition. Using such decompositions in a hierarchical ....

....projection tests (about 200 arithmetic operations) as proved by the authors through their separating axis theorem. Routines for building OBB trees, as well as for performing fast overlap tests between them can be found in the RAPID interference detection package [35] OBBs have also been used in [31] where, if interference between boxes is not discarded, OBBs are further subdivided into voxels and, if needed, the interference test is finally applied to boundary features of objects. On the other hand, a hierarchy based on axis aligned bounding boxes (AABB) has the advantage that the ....

Garc' ia-Alonso, A., Serrano, N., and Flaquer, J. Solving the collision detection problem. IEEE Computer Graphics and Applications 14, 3 (May 1994), 36--43.

.... reconstruct new images from the reprojected pixels of reference images [14] The bounding box heuristic is also common in algorithms for modeling, including techniques that de ne complex shapes as Boolean combinations of simpler shapes [10] or that verify the clearance of parts in an assembly [6]. Animation algorithms also exploit bounding boxes, especially collision detection algorithms for path planning [12] and the simulation of physically based motion [1, 11, 15] While there is ample empirical evidence that the bounding box heuristic improves performance in practice, the goal of ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36-43, 1995. 10

.... reconstruct new images from the reprojected pixels of reference images [36] The bounding box heuristic is also common in algorithms for modeling, including techniques that de ne complex shapes as Boolean combinations of simpler shapes [26] or that verify the clearance of parts in an assembly [22]. Animation algorithms also exploit bounding boxes, especially collision detection algorithms for path planning [32] and the simulation of physically based motion [11, 31, 38] 16 3.1.2 The Bounding Box Heuristic in Collision Detection The use of bounding boxes in collision detection is ....

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14:36-43, 1995.

....depicts a few representative runs of our experiments, based on two methods: the mesh based method, and the bounding volume hierarchy method. Details of our algorithms, implementation issues, and experimental results are reported in [3, 4] For other recent related work on collision detection, see [1, 2, 6, 7] 2 Mesh Based Method This method is motivated by recent results on efficient ray segment shooting queries using meshes (triangulations) of low stabbing number [5, 8] The complexity of an intersection query with an object F is proportional to the number of cells (tetrahedra) of the mesh that are ....

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14:36--43, May 1994.

.... where the motion is subject to dynamic constraints or external forces and cannot typically be Also with Computer Science Department, University of North Carolina, Chapel Hill, NC 27599 3175; lin cs.unc.edu 1 expressed as a closed form function of time, various algorithms have been proposed [1,2,4,5,7,12]. At the same time, the emphasis in the computational geometry has been on the theoretically efficient intersection detection algorithms [10] Most of them are restricted to a class of static objects and non trivial to implement. Different methods have also been proposed to overcome the bottleneck ....

Garcia-Alonso, A., Serrano, N., and Flaquer, J., Solving the Collision Detection Problem, IEEE Computer Graphics and Applications, 13 (3): pp. 36-43, 1994.

....have been widely used to design efficient algorithms. Typical hierarchies include k d trees and octrees, R trees and their variants, cone trees, BSPs [NAT90] and their extensions to multi space partitions [WG91] and spatial representations based on space time bounds or four dimensional tests [AANJ94, Hub93] The set of BVs for BVHs include spheres [Hub93, Qui94] axis aligned bounding boxes [BKSS90, HKM95] oriented bounding boxes [GLM96, BCG 96] approximation hierarchies based on S bounds [Cam91] spherical shells [KPLM98] and k DOPs [HKM96, KHM 98] There is also literature on ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

....taking a courser grid, but in that case the list of a very large cell will include environment triangles which aren t even nearby the moving object. For the models they tested, they found the best grid resolution ranged between 5 Theta 5 Theta 5 and 50 Theta 50 Theta 50. Garcia Alonso et al. [Alonso94] also use a rectilinear grid to eliminate some of the polygon overlap tests. Each model is covered by a bounding box (which is oriented with the model) which is further subdivided into a regular grid of predetermined resolution. Grid cells hold a list of polygons which touch the cell. For a given ....

A. Garcia-Alonso, N Serrano, J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, Vol. 14, pp36-43, 1994.

....the pair processing weakness for polyhedral objects. An advantage of BSP trees is that for rigid objects, the structure of a tree can be optimized to improve performance; Naylor [19] describes several optimization techniques. To accelerate their pair processing algorithm, Garcia Alonso et al. [9] associate a precomputed regular grid with each object; as objects rotate, however, the grids lose some of their efficiency. Sclaroff and Pentland [24] present a pair processing algorithm that approximates surfaces with deformed superquadrics. The authors do not demonstrate that their approach ....

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14(3):36--43, May 1995.

....structures include cone trees, k d trees and octrees [Sam89] sphere trees [Hub93, Qui94] trees based on S bounds [Cam91] etc. Binary space partitions (BSP) NAT90] and extensions to multi space partitions [BV91] and spatial partitionings based on space time bounds or four dimensional testing [AANJ94, Cam90,Can86, Hub93] have been used. All of these hierarchical methods do very well in performing rejection tests whenever two objects are far apart. However, when the two objects are in close proximity and can have multiple contacts, these algorithms either use subdivision techniques or check very large number of ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994. Ming C. Lin & Stefan Gottschalk 15

....include cone trees, k d trees and octrees [Sam89] sphere trees [Hub93, Qui94] trees based on S bounds [Cam91] etc. Other spatial representations are based on BSP s and its extensions to multi space partitions [WG91] spatial representations based on space time bounds or four dimensional testing [AANJ94, Cam90, Can86, Hub93] and many more. All of these hierarchical methods do very well in performing rejection tests , whenever two objects are far apart. However, when the two objects are in close proximity and can have multiple contacts, these algorithms either use subdivision techniques or check ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

....in [8] runs in roughly constant time even when the closest features change. However, they are not applicable in the case of a surgery simulator, since organs are generally non convex, and deform over time. Among the collision detection methods that are applicable to more general polygonal models [10, 2, 12, 4, 11, 13, 5, 7], almost all of the optimizations rely on a pre computed hierarchy of bounding volumes. The solutions range from axis aligned box trees, sphere trees [12, 11, 7] or BSP trees, to more specific data structures [2] All these techniques, which perform very efficient rejection tests, may ....

....by using hierarchies of oriented bounding boxes instead of axis aligned boxes. Section 5 will compare our new method with the public domain software package RAPID that implements this technique. A last issue in collision detection is the ability to perform dynamic rather than static detection [10, 4]: moving objects may interpenetrate between consecutive time steps, so the intersections should be computed between the 4D volumes that represent the solids trajectories during a time step rather than between static instances of the solids. In the context of a large environment with lots of ....

[Article contains additional citation context not shown here]

A. Garica-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

.... R trees and their variants [BKSS90] trees based on S bounds [Cam91] strip trees, and boxtrees [ZF95] Other spatial representations are based on BSP s [NAT90] and its extensions to multi space partitions [WG91] spatial representations based on space time bounds or four dimensional testing [AANJ94, Cam90, Can86, Hub94] and many more. All of these hierarchical methods do very well in performing rejection tests , whenever two objects are far apart. However, when the two objects are in close proximity and can have multiple contacts, these algorithms either use subdivision techniques or check very large number of ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

....trees and octrees [31] sphere trees [20, 28] R trees and their variants [5] trees based on S bounds [7] etc. Other spatial representations are based on BSP s [24] and its extensions to multi space partitions [34] spatial representations based on space time bounds or four dimensional testing [1, 6, 8, 20] and many more. All of these hierarchical methods do very well in performing rejection tests , whenever two objects are far apart. However, when the two objects are in close proximity and can have multiple contacts, these algorithms either use subdivision techniques or check very large number of ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

.... that reconstruct new images from the reprojected pixels of reference images [23] Bounding boxes are also common in algorithms for modeling, from techniques that define complex shapes as Boolean combinations of simpler shapes [18] to techniques that verify the clearance of parts in an assembly [11]. Animation algorithms also exploit bounding boxes, especially collision detection algorithms for path planning [21] and the simulation of physically based motion [2, 20, 25] While empirical evidence demonstrates that the bounding box heuristic improves performance in practice, the goal of ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36--43, 1995.

....shapes. One of the earliest examples of this idea is the work of M antyl a and Tamminen [30] Kitamura et al. 26] present a more modern variation that uses octrees. Three recent algorithms with narrow phases that use subdivision techniques the work of Smith et al. 44] Garcia Alonso et al. [14] and Ponamgi et al. 36] deserve particular attention. Although their broad phases create the temporal inaccuracy discussed in Section 2, these algorithms have the advantage that they achieve real time performance for some challenging situations. In the Smith et al. algorithm, the narrow phase ....

....at the first collision) Few real applications end at the first collision, however. It is unclear how the algorithm would perform after collision response, which can cause multiple objects to collect in colliding and nearly colliding configurations for many time steps. The Garcia Alonso et al. [14] algorithm precomputes a one level grid subdivision for each object. The narrow phase uses these grids to search for intersecting faces within the overlap of object bounding boxes. When detecting interference in an unfolding satellite antenna (with 50 jointed objects and 1500 total faces) the ....

Alejandro Garcia-Alonso, Nicol'as Serrano, and Juan Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14(3):36--43, May 1995.

....first and than a static interference test is performed at each time. There are several methods how to improve the efficiency of this method. They employ objects bounds, a time step varying, the caching of data and the reduction of number of necessary tests. Garcia Alonso et al. introduces in [GASF94] an interest matrix, containers, boundaries, and voxel spaces as an approximation for fast estimation of nonexisting collisions. Pobil and Serna in [PS95] elect interior and exterior sets of spheres for the approximation since the sphere is invariant under translation and rotation ....

....FS is the slowest and non efficient algorithm it is still used for its robustness and simplicity. The complexity is rather high (in comparison with other algorithms) but it is stable and predictable during the simulation and can be lowered by speed up techniques like a space partitioning in [GASF94] or spacetime bounds in [Hub95] 5.2 Binary Search Binary search (BS) was developed as an attempt to speed up the fixed step algorithm. The idea is simple, knowing a time point t out , the subject is outside the observer s range, and t in , it it inside, the place of approximately accurate ....

[Article contains additional citation context not shown here]

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, pages 37--43, May 1994.

.... reconstruct new images from the reprojected pixels of reference images [14] The bounding box heuristic is also common in algorithms for modeling, including techniques that define complex shapes as Boolean combinations of simpler shapes [10] or that verify the clearance of parts in an assembly [6]. Research supported by NSF grants 9628190 and 9501494. y Department of Computer Science, Washington University, St. Louis, MO 63130. Email: yzhou cs.wustl.edu. z Department of Computer Science, Washington University, St. Louis, MO 63130. Email: suri cs.wustl.edu. Animation algorithms ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36--43, 1995.

....is particularly trivial) and then tracking the closest pairs of points between all pairs of convex pieces. This method is highly effective as long as there are not too many convex pieces involved; it has the advantage of not relying on any of the objects remaining stationary. Other work includes [13] and the thesis work of Hubbard [17, 18] who uses approximations of objects (covering by disks) to speed up collision detection. Very recent papers include [19] and [27] An Application to VR in Aircraft Manufacturing We were drawn into this line of research when the Boeing VR group came to us ....

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14:36--43, May 1994.

No context found.

A. Garca-Alonso, N. Serrano, J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, pp. 36--43, 1994. 2

....Shaffer et al. 4] have used an octree to approximate objects. Good results have been achieved for convex polyhedra [16, 3, 8] Almost all approaches for general, non convex objects utilize some sort of BV hierarchy. A few nonhierarchical approaches have been taken as well. GarciaAlonso et al. [7] partition the set of polygons of an object by a uniform grid. Sphere trees were developed by [13, 14, 19] Gottschalk et al. 11] developed a bounding box hierarchy based on oriented boxes (OBBs) Its box overlap test is much more expensive than for DOPs. Aligned Box Trees have been presented in ....

A. Garca-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, pages 36--43, May 1994.

....consider the issue of finding the exact time of primal contact between two polyhedra. The BoxTree data structure is a binary tree, which is a non uniform, adaptive space subdivision. Because it is non uniform, we suspect it to out perform approaches using uniform subdivisions, like grids (as in [9]) Furthermore, since it is adaptive, it improves the expected runtime performance. The leaves of a BoxTree contain edges and polygons which define the associated polyhedron. A tree construction algorithm tries to build an optimal tree with respect to the collision detection algorithm. The ....

....can be solved more efficiently than the construction problem [7, 6, 27] In the field of robotics, a completely different approach has been pursued: collisions are detected in configuration space (see [8] for example) This approach seems to be well suited for path planning. As stated earlier [9], the representation of objects has great impact on collision detection algorithms. Non b rep representations, e.g. octree, BSP, CSG, etc. allow need quite different approaches [4, 23, 24, 30] For collision avoidance systems, an approximate collision detection is quite appropriate [5] 15] ....

[Article contains additional citation context not shown here]

A. Garc ' ia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, ?(?):36--43, May 1994.

....in [8] runs in roughly constant time even when the closest features change. However, they are not applicable in the case of a surgery simulator, since organs are generally non convex, and deform over time. Among the collision detection methods that are applicable to more general polygonal models [10, 2, 12, 4, 11, 13, 5, 7], almost all of the optimizations rely on a pre computed hierarchy of bounding volumes. The solutions range from axis aligned box trees, sphere trees [12, 11, 7] or BSP trees, to more specific data structures [2] All these techniques, which perform very efficient rejection tests, may ....

....by using hierarchies of oriented bounding boxes instead of axis aligned boxes. Section 5 will compare our new method with the public domain software package RAPID that implements this technique. A last issue in collision detection is the ability to perform dynamic rather than static detection [10, 4]: moving objects may interpenetrate between consecutive time steps, so the intersections should be computed between the 4D volumes that represent the solids trajectories during a time step rather than between static instances of the solids. In the context of a large environment with lots of ....

[Article contains additional citation context not shown here]

A. Garica-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

.... such as the visiblesurface determination [2] the view frustum culling [5] and the reprojected pixel imaging [12] Many modeling algorithms also use bounding boxes for defining complex shapes as boolean combinations of simpler shapes [8] or for verifying the clearance of parts in an assembly [3]. Animation algorithms use bounding boxes for collision detection in path planning [10] or simulation of physically based motion [1, 9, 13] The use of bounding boxes in collision detection is representative of its use in other algorithms, so we will focus on collision detection. Many collision ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36--43, 1995.

.... such as the visible surface determination [3] the view frustum culling [5] and the reprojected pixel imaging [10] Many modeling algorithms also use bounding boxes for de ning complex shapes as boolean combinations of simpler shapes [7] or for verifying the clearance of parts in an assembly [4]. Animation algorithms use bounding boxes for collision detection in path planning [9] or simulation of physically based motion [1, 8, 11] The use of bounding boxes in collision detection is representative of its use in other algorithms, so we will focus on collision detection. Many collision ....

A. Garcia-Alonso, N. Serrano and J. Flaquer. Solving the Collision Detection Problem. IEEE Computer Graphics and Applications, 14, pp. 36-43, 1995.

....hierarchical bounding box scheme. Objects are surrounded by bounding boxes. If the bounding boxes overlap, indicating the objects are near each other, a more precise collision test is applied. Approximate collision detection schemes using a series bounding boxes, spheres, or voxels can be found in [AANJ94, HBZ90]. More recently, Hubbard [Hub93] has formulated an approach where accuracy is traded for speed. The approximation schemes can be used to deal with general non convex objects. A non convex object can be broken into these sub units which are tested for collisions. However, to achieve desired ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

....trees and octrees [31] sphere trees [20, 30] R trees and their variants [5] trees based on S bounds [7] etc. Other spatial representations are based on BSP s [27] and its extensions to multi space partitions [32] spatial representations based on space time bounds or four dimensional testing [1, 6, 9, 20] and many more. All of these hierarchical methods do very well in performing rejection tests , whenever two objects are far apart. However, when the two objects are in close proximity and can have multiple contacts, these algorithms either use subdivision techniques or check very large number of ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

....of Rolhf and Helman [19] and the human figure animation of Granieri et al. 9] 2 Why Time Critical Collision Detection The idea of trading accuracy for speed in collision detection may seem controversial. Recent traditional (non time critical) detection algorithms by Garcia Alonso et al. [6], Ponamgi et al. 17] and Smith et al. 20] do provide real time performance for some challenging test cases. Nevertheless, there are several arguments that a time critical detection algorithm is the most promising way to meet the current and future needs of virtual environments. A central ....

....elsewhere [12] confirm that accuracy does, in fact, increase with processing time. To put the accuracy of a time critical algorithm in perspective, note that few traditional algorithms actually provide full accuracy. Most traditional algorithms including the three fastest from the literature [6, 17, 20], mentioned above allow temporal inaccuracy by detecting collisions only once per frame. This temporal inaccuracy creates spatial inaccuracy equal to the distance an object travels between frames. A human figure walking at 5 miles per hour, for example, travels 8.8 inches per frame, given 10 ....

A. Garcia-Alonso, N. Serrano, and J. Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 14(3):36--43, May 1995.

.... goal in robotics has been the planning of collision free paths between obstacles [15] This differs from virtual environments and physicallybased simulations, where the motion is subject to dynamic constraints or external forces and cannot typically be expressed as a closed form function of time [1, 3, 11, 18, 20, 21]. At the same time, the emphasis in the computational geometry has been on theoretically efficient intersection detection algorithms [22] Most of them are restricted to a static instance of the problem and are non trivial to implement. For convex 3 polytopes 1 linear time algorithms based on ....

....convex 3 polytopes 1 linear time algorithms based on linear programming and tracking closest points [10] have been proposed. More recently, temporal and geometric coherence have been used to devise algorithms based on checking local features of pairs of convex 3 polytopes [3, 17] Alonso et al.[1] use bounding boxes and spatial partitioning to test all O(n 2 ) pairs of arbitrary polyhedral objects. Different methods have been proposed to overcome the bottleneck of O(n 2 ) pairwise tests in an environment of n bodies. The simplest of these are based on spatial subdivision. The space is ....

A.Garica-Alonso, N.Serrano, and J.Flaquer. Solving the collision detection problem. IEEE Computer Graphics and Applications, 13(3):36--43, 1994.

Online articles have much greater impact   More about CiteSeer.IST   Add search form to your site   Submit documents   Feedback  

CiteSeer.IST - Copyright Penn State and NEC