IJCSI International Journal of Computer Science Issues, Vol.

7, Issue 1, 2010 1
ISSN (Online): 1694-0784
ISSN (Print): 1694-0814

New Approaches towards the Collision detection

Manish Kumar1, Gaurav Mishra2 , Kamal Kumar Srivastava3and Ashish Avasthi4
1
Uttar Pradesh Technical University, CS Dept.., Bansal Institute of Technology and Management
Lucknow, Uttar Pradesh, India, [email protected]
2
Uttar Pradesh Technical University, MCA Dept.., Shri Ramswaroop Memorial College of Engg. & Mgmt.
Lucknow, Uttar Pradesh, India,[email protected]
3
Uttar Pradesh Technical University CS., Shri Ramswaroop Memorial College of Engg. & Mgmt
Lucknow, Uttar Pradesh, India,[email protected]
4
Uttar Pradesh Technical University, MCA Dept.., Shri Ramswaroop Memorial College of Engg. & Mgmt
Lucknow, Uttar Pradesh, India,[email protected]

Abstract During several years, people work on different collision

detection approaches but recently, few papers appear dealing
Collision detection on industrial digital mock-up is one of the with a new type of problem: speeding up the detection using
challenging problems for real-time interaction in virtual reality hardware power (CPU and GPU) [19, 48].
applications. The fast increase of graphics hardware In the wide range of collision detection algorithms we can notice
performance, multiplication of cores number and recent that recent articles [4, 22] have a low complexity and provide
improvements on their programmability, bring new directions to perfect collision detection. But used with million and million of
the optimization of collision detection algorithms. objects in a huge environment, they are inefficient for a real-time
Since few years methods appear handling General-Purpose interaction. We may hope a constant evolution of processors
Processing on Graphics Processing Unit (GPGPU) and more power to resolve the real-time problem but trend is no more on
recently using multi-cores. In this paper we present an analysis that but rather in the processors multiplication. Hardware
on new trends in collision detection performance and we study graphics is also subjected to an impressive power evolution. So it
the use from GPU to grid for virtual reality applications. Here we appears that having a different look on the real- time collision
not only deal with algorithmic improvement but also propose a detection problem centered on hardware performance can not be
first approach on the new link-up between two fields of computer ignored.
graphics namely virtual reality (collision detection) and In the following section 2 gives a brief survey on collision
computers performance. Further we will also discuss how to use detection algorithms. In section 3 we describe the hardware and
of fractal theory in understanding the collision detection by architecture evolution, followed by software and middleware
studying comparatively less developed network. evolution in section 4. We present in section 5, a survey on the
new link-up between collision detection and architecture
performance. At the end we expose our personal point of view on
Keywords: Collision detection, performance, parallel
this new link-up.
computing, scalable, graphics hardware, Fractals, GPGPU,
Since in this paper we also presented fractal based algorithm so a
clusters, grids
little intro of fractal is also required. The name ‘Fractal’ came
from a Latin word ‘fractus’, which means’ to break Indeed the
1. Introduction word fractal, firstly given by B.B.Mandelbort in 1975, who is
also called father of fractal geometry.
Industrial applications of virtual reality (VR) become more and In the words of Mandelbrot,”Fractal is a rough fragmented
more sized and the performance level for a real-time interaction geometric shape that can be subdivided in parts each of which is,
of users is no longer satisfying. From more than thirty years, at least approximately, a reduced copy of the whole”[6-7].
collision detection became the principal bottleneck of real-time Fractal is a geometric pattern that is repeated at ever smaller state
VR applications. Collision detection (CD) is a wide field dealing to produce irregular shapes and surfaces that cannot be
with, apparently, an easy problem: determine if two (or several) represented by classical geometry. Fractals are used especially in
objects collide. It is used in several domains namely physically- computer modeling of irregular patterns and structures in nature
based simulation, computer animation, robotics, mechanical [54 - 56].
simulations (medical, biology, cars industry…), haptics
applications and video games. In all of these applications, real-
time performance, efficiency and robustness are more and more
required.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 2
www.IJCSI.org

intersect. Beginning with this assumption, different methods

have been proposed to compute this separating plan. Some of
them use a vectorial product exploiting temporal coherence or
compute a bounded distance between objects to build a
separating plan [25]. Other methods proposed to determine inter-
objects distance. The most famous one is the GJK algorithm [24]
that uses Minkowski difference on polyhedrons. A lot of GJK
algorithm improvements have been proposed:
- ‘Enhancing GJK’ [72] uses hill climbing method to optimize
computing time.
- ‘ISA-GJK’ [27] uses data caching to increase performances and
uses a faster method to build separating plan using temporal
Figure 1. Mandelbrot Fractals coherence.
- MS [17] couples GJK algorithm with Lin-Canny algorithm.
2. Collision detection Lin-Canny approach [51] or Voronoï Marching was the first
algorithm working with objects primitives. Space around objects
We now expose a short survey on the collision detection, for is divided in Voronoï regions that allow detecting closest
more details we refer to excellent surveys on the topic [64, 68, features pairs between polyhedrons. The V-Clip algorithm [3]
51, 50]. Collision detection has generated a wide range of operates on a pair of polyhedron, defining closest points between
problems families: convex or non-convex objects, 2-Body or N- pairs in terms of closest features of the polyhedron. Another way
Body simulations, rigid or deformable objects, continual or to compute distance between objects is to perform pre-
discrete methods. Algorithms are also linked to geometric used computations on polyhedrons in order to reduce intersection
models (polygonal, Constructive Solid Geometry (CSG), implicit detection complexity.
or parametric functions). All of these problems reveal the high It is also possible to compute the interpenetration between two
complexity and difficulty of this field of study. objects. Several approaches use GJK algorithm to compute
Given n moving objects in a virtual environment, testing all penetration depth [72]. Gregory et al. [1] propose to extend
objects pairs tend to perform O (n²) pair wise checks. When n is Voronoï marching method coupling with temporal
large it becomes a computational bottleneck. Collision detection coherence. Given direction, depth penetration can be compute
is, since Hubbard [65], represented and built as a pipeline. This
with Dobkin’s hierarchy [22]. The use of normales space to find
one is composed by three parts namely, broad-phase, narrow-
translation direction has been proposed [81]. More recently, a
phase and exact-phase (core-phase). The goal of this pipeline is
method detects interpenetration with ray casting [23].
to apply successive filters in order to break down the O(n²)
complexity. These filters provide an increasing efficiency and Detection between non-convex polyhedrons
robustness during the pipeline traversal. As input, the pipeline
takes all the geometric data of the simulation and feeds in output Algorithms described previously are suited for convex
the collision response module. Nowadays pipeline is shared in polyhedrons, use it with non-convex objects and they would do
two phases: broad and narrow phases (exact phase being mistakes on the non-collision detection or collision non-
included in narrow-phase). The first part of the pipeline, called detection. To avoid expensive computing time, most of
the broad-phase, is in charge of a quick and efficient removal of approaches use a bounding volume hierarchy that allows creating
the objects pairs that are apparently not in collision. In the other series converging to minimal distance [18]. Quinlan [69] uses
hand, narrow-phase is in charge of determining exact collision spheres to prune parts of his models, Sato et al. [82] propose to
detection. couple Quinlan’s spheres-trees with GJK algorithm. Other
methods propose to find a separating plan with points of one
2.1 Exact collision detection object in the negative side and points of the other in the positive
side [30]. It is also possible to test objects primitives in order to
The narrow phase constitutes the pair wise tests within subgroups determine if a vertices of an object is inside another one or if a
of potentially colliding objects. We present important parts of segment intersects an object face [37]. Comparisons on objects
exact collision detection starting by basic tests between triangles [53] or rectangles are possible.
polyhedrons, following with bounding volumes and their To compute penetration depth between non-convex objects,
hierarchies. Fisher et al. [73] show that using Minkowski sum can have an O
2.1.1 Basic tests (n6) complexity. Dobkin et al. [22] propose a method to compute
this penetration between a non-convex object and a convex one,
In case of polygonal representations, tests are made on objects but with two non-convex objects, complexity becomes too high.
primitives. With CSG representations, tests are made on A trivial approach consists in separating object into convex parts
elementary geometrical objects (cubes, cones…). Implicit and computing penetration between overlapping convex parts
function or parametric surface representations involve tests with [81]. Distance fields can also be used for the penetration depth
membership function. We present different algorithms used to computing [73].
detect collision with polygonal representation.
2.1.2 Bounding Volume
Detection between convex polyhedrons
Most of strategies use bounding volume hierarchies to perform
It is easy to imagine that two objects separated by a plan don’t collision tests. There are a lot of bounding volume suchas
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 3
www.IJCSI.org

sphere[65],Axis- Aligned-Bounding-Box AABBs [27], Oriented-

Bounding-Box OBBs [74], discrete oriented polytopes (k-DOP)
[30] or convex hulls. As explained in [10], many other types of
volume have been suggested as bounding volume namely,
sphere-swept volumes, cones, cylinders, spherical shells,
ellipsoids and zonotopes. Using bounding volume (BV) to
perform tests before testing the object itself, highly improves
performance. Although tests have been simplified, to test
collision between two objects, the same pairwise check is still
performed. Bounding volume hierarchies (BVH) allow arranging
BV into a tree hierarchy in order to reduce the number of tests to
perform. An excellent description on these BVH and a Figure 2: “Sweep and Prune” with oriented-bounding boxes.
comparison between their performance can be found in [10, 68].
Construction of these trees is performed according to three 3. Hardware and Architecture evolution
primary categories: top-down, bottom-up and insertion.
Deformable objects are very challenging for BVH because We now expose the evolution of computer hardware and
hierarchy structures (trees) have to be updated when an object architecture that can be used to improve collision detection. First
deforms itself. As trees can not be updated at each time step, we start with a presentation of graphics hardware (GPU) that has
better solutions are compulsory [27, 49]. been highly improved those last years. We then present evolution
of CPU from simple-core to multi core and many-core. We
2.2 Accelerative steps introduce several features of the architecture evolution in order
to take care of it during collision detection optimization. At last,
Broad-phase algorithms are classified into three main families we briefly present the cluster and grid architectures, their use and
[68], namely brute force method with bounding volumes, spatial different problems that can be encountered in performing real-
hashing and topological and cinematic methods. time collision detection.
2.2.1 Brute force
Brute force approach is based on comparing bounding volumes
of the overall objects pairs to determine if they are in collision or 3.1 GPU evolution
not. This test is very exhaustive because of its O(n²) pairwise
Recent years have seen the evolution of graphics hardware from
checks.
fixed function units toward an increasingly programmable
2.2.2 Spatial partitioning graphics pipeline. Contrary to CPU, Graphics Processing Unit
(GPU) has a very important power evolution since few years (cf.
Spatial hashing method is based on a simple rule: two objects Figure 3). This impressive evolution can be explained by the way
situated in distant space sides have no chance to collide during that in one hand, CPU is a generalist processor (cf. Figure 5) that
next time step. To divide space into unit cells several methods deals with ordinary data expressing a high level of dependencies,
have been proposed: regular grid [65], octree [71], quad-tree, several of its components are in charge of the data stream control
Binary Space Partitioning (BSP), k-d tree structure [44] or while its memory latency period is hidden by data caching. In the
voxels. Subdivisions made on space can be independent from the other hand GPU processors are well-suited to highly
environment [44] and can also be given by it. This technique is parallelizable computations (cf. Figure 6a). It deals with
only accurate when the environment is static (2-body context). It independent data so it does not need a sophisticated data stream
is also possible to use tetrahedron meshing or use a constraints control and its memory latency period is hidden by
set projected on a high-dimensional space (six dimensions) [38]. computations. For instance, we compare principal graphics cards
Teschner et al. [49] employ a hash function to compress a of ATI1 and Nvidia2 in 2008 (cf. Figure 3).
potentially infinite regular spatial grid.
2.2.3 Topology and cinematic
Topology methods are based on the positions of objects in
relation to others. A couple of objects that are too far one to the
other is deleted. One of the most famous methods is called the
“Sweep and Prune” [39] approach and consists in projecting
objects coordinates on axis. If the projection reveals an
overlapping of objects coordinates, they are probably in collision
and they are then given to narrow-phase. This method is, in
general, used with bounding volume like AABBs [39] or OBBs Figure 3: Comparison of recent graphics cards of Nvidia and ATI
[70] (Figure 2).
On the contrary, the cinematic approach takes care of the objects GPU can be imaged (in the GeForce 8800 version) as a big
movement, if objects are moving away, they can not collide. computing unit with 16 cells of 8 ALUs (Arithmetic Logical
Vanecek [29] used cinematic of the objects and back-face culling Unit) handling each one 4 threads. These 16 cells are able to
manipulate all current instructions on 512 threads, with a flow of
technique to speed up collision detection.
256 operations per cycle. GeForce 8800 has been described as a
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 4
www.IJCSI.org

GPU equipped with 128 processors allocated to 8 high frequency contributions are not only in the algorithms improvement but
partitions (1350 MHz). also in the algorithms architecture modification. As we can not
Bandwidth on GPU is also higher than on CPU [43] but a hope a continual evolution of processors we have now to study
fundamental problem of performing computations on GPU is the what it is possible to do with multi-core in collision detection
bandwidth between CPU and GPU (only 4 GB/s). A recent algorithms.
solution* proposes to hide data-transferring time by using Now a days it is impossible to present CPU without dealing with
concurrent memory copy (between CPU and GPU) and central memory handling. Indeed, on a multi or many cores,
execution on GPU. there is a very complex cache handling between cores. This
handling is continuously improved to increase computer
performance. Cache and memory handling is Cache and memory
handling is another point that cannot be ignored in the
optimizations of the collision detection performance.

3.3 Clusters and Supercomputers

A cluster, imaged as a computer grape, is a powerful engine with
high performance that can be useful for real-time collision
detection. It is composed by localized machines connected
through a local network (for instance Ethernet or Giga-Ethernet).
In comparison to distributed units in GPU and multi or many
cores in CPU, a cluster, with a higher scale, can be seen as a
many or multi computers machine (cf. Figure 5c). A survey [5]
has been done on different approaches that have been developed
Figure 3: Comparison of the CPU and GPU evolution from 2003 to 2007 to use PC clusters for virtual reality applications. This survey
also presents middle-ware allowing using maximum of cluster
3.2 CPU evolution performances (section 4).
Compared to actual outlook, it seems clear that Gordon Moore Differences between a cluster and a supercomputer become very
was a lucky man. Since 1965, he predicts a duplication of the thin because they use same CPUs and GPUs connected with a
number of transistors on a microprocessor each two years. high performance network working on the same Operating
During more than forty years, this guesswork seems exact but we System. Contrary to personal computers, clusters and
know now that physical limits (power and heat) prevent this supercomputers receive great attention on their communication
duplication. architecture. The problem changes when you have to manage 10,
50, 500 or 2000 nodes. Several network topologies have been
proposed, for instance: cubes, hypercube or torus.
3.4 Grids
A grid computing is an infrastructure composed by a mass of
non-homogenous informatics resources (PC-clusters, computers,
servers, mobiles…). Grid sites are geographically separated and
it is usually used for huge scientific computations (cf. Figure 5d).
A grid is a bit different than clusters and computers because it is
not in the parallel computing world but rather in the distributed
world. A middleware is integrated into grids in order to abstract
Figure 4: Comparison of the architecture of CPU and GPU
all resources and so to exploit computation power (processor,
memory…). Using a grid can reveal some more important
Now a days trend tends to be duplication of cores (cf. Figure 5b)
problems than on clusters because of the resources volatility, the
in computers and parallel architectures. The first personal
access rights and the latency time. Communication between grid
computer with a core-duo arrived in 2005 with AMD1 followed
resources is done with optical fiber and every one knows that its
by Intel3. In 2006 Sun4 presented its new octo-core called
bandwidth physica limit is light speed. But as optical fibers go
Niagara2. Intel presents last year a 32 in-order x86 cores [46]
through signals amplifiers, bandwidth is reduced. This speed
and Sun recently announce 80 cores computer. Another emerging
reduction has to be taking account into distributed computation
CPU concept is many-core: the computer dynamically adapts the
for real-time interaction. Algorithms that would work on grid
number of active cores with respect to the user needs. Many-core
have to be more predictive than detective due to the latency.
is useful because when people do not need the entire power of
These algorithms have also to provide a dynamic resources
cores, computer turns off some of them. Until now, 3D objects
management because of their volatility
and virtual environments grew up in parallel to processor power,
so researchers were continuously looking for improvements on 3.5 Conclusion
the collision detection algorithms in order to increase their
precision, robustness and efficiency [21, 23]. But now, For any architectural levels from GPU to grid, we can notice a
processors power stays roughly constant while virtual fractal-like hierarchy (cf. Figure 5). In this figure, even if
environments are more and more sized, so new scientific architectures schemes have been obviously simplified, they are
still representative.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 5
www.IJCSI.org

4. Collision detection algorithm

Recent years have seen an increasing interest of performing
collision detection algorithms taking into account computer
architecture. We present three main families of architecture-
based algorithms: GPU, Multi-threads and Multi-cores.
4.1 Existing Algorithm

4.1.1 GPU Based Algorithm : Image-based algorithms have

been proposed to exploit the growing computational power of
graphics hardware. GPU is very efficient in rasterisation of
polygons; GPU-based collision detection algorithms rasterise the
objects and perform either 2D or 2.5-D overlap tests in screen
space [63]. Given several objects meshes, it returns pairs of
objects primitives that are then computed on the CPU. A good
advantage of using graphics hardware is the un-use of
precomputed volumetric data structures and its use with rigid or
deformable objects. GPU can also be useful to compute distance
fields using a uniform spatial grid [2].
Furthermore, visibility computations can be performed using
occlusion queries and used to compute both intra- and inter-
object collisions among multiple objects [63]. A technique using
image-space have been proposed and compared to a CPU-based
implementation; results show that GPU accelerates collision
detection in complex environments but CPU-based methods
provides more flexibility and better performance in small
environments [4]. Broad phase is also made with GPU using
image-space visibility queries [63].
Cinder [20] is an algorithm exploiting GPU to implement a ray-
casting method to detect collision. When a ray strikes an edge, a
count of the difference in the number of back-facing and front
facing polygons lying between the edge point and the ray’s
origin at the view port is made. GPU-based algorithms for self-
collision and cloth animation have also been introduced by
Govindaraju et al. [64]. An efficient backward voxel-based
AABB hierarchy method was proposed to handle deformable
surfaces that are highly compressed using graphics hardware
[26].
4.1.2 Multi-threads based algorithm : Since few years,
researchers are working on the implementation of multithreaded
algorithms in collision detection and more precisely in dynamics
molecular simulation. Lewis et al. [47] propose a new
multithreaded algorithm to simulate planetary rings. An
evaluation of the performance of a parallelized back-end of the
pipeline has been made by Zachmann [30] and shows that if the
environment density is large compared to the number of
processors, then good speed-ups can be noticed. This evaluation
did not parallelize the other phases of the pipeline.

4.2 New Algorithm

4.2.1 Multi-cores based algorithm : Very recently, few

papers appear dealing with new parallel collision detection
Figure 5: Comparison of simplified architectures of GPU, quad-core, PC- algorithms using multi cores. Tang & al. [48] propose to use a
Cluster and grid hierarchical representation to accelerate collision detection
queries and an incremental algorithm exploiting temporal
In all of them, there are several computational units and a lot of coherence the overall is distributed among multiple cores. They
memories and caches units through communication architectures. obtained a 4X-6X speed-up on a 8-cores based on several
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 6
www.IJCSI.org

deformable models. Kim & al [19] propose to use a feature- These features of fractal i.e. self similarity and iterative
based bounding volume hierarchy (BVH) to improve the procedure for generating recursive fractals (Figure 6) are used in
performance of continuous collision detection. algorithm. With these properties we can detect the errors in a
small portion of the object and then later on spread it with the
They also propose novel task decomposition methods for their
help of a complex mathematical equation using Picard and
BVH-based collision detection and dynamic task assignment
Superior iterations [54 - 58]. By the means of this fractal based
methods. They obtained a 7X-8X speed-up using a 8-cores
algorithm one can detect pattern of advance stages and have a
compared to a single-core.
say in error detection. Up to this stage basic problems was to
4.2.2 Fractals based algorithm : The basic concept of examine the large networks as a whole or simultaneously but
fractals is that they contain a large degree of self similarity. This with the advent of this algorithm things can be examine in a
means they usually contain little copies of themselves buried small part at first go and later on can be seen in completion.
deep within the original and also have infinite details. The
repetition of form over a variety of scales is called self 5. Revisited CD pipeline (Effect of the new
similarities. A fractal looks similar to itself on a variety of scales.
The bigger eddies in a turbulent flow look much the same as the
algorithms)
smaller ones, and vice versa. Through this survey, it appears that the architecture of collision
The backbone of the fractal is iteration method i.e. feedback detection algorithms needs to be improved to face real-time
system. In other words the same process is performed repeatedly. interaction. In this way, we are thinking about reviews of the
The output of the system acts as an input for next one. One of the collision detection pipeline. We work on a global parallelization
unique things about fractals is that they have non integer of the pipeline in order to avoid starvation of each phase. They
dimensions. That is why we are in the 3rd dimension, looking at would not stop and run continuously during simulation. Broad-
this on a flat screen which can be considered more or less the phase is still pruning un-colliding objects pairs and feeds an
2nd dimension; fractals are in between the dimensions. Fractals object pairs buffer that is, in the same time, used by narrow-
come into two major categories i.e. Deterministic Fractal And phase to provide exact tests. This pipeline can be imagined as a
random Fractal. The first categories consist those Fractal that are double-buffer with access control. This parallelized pipeline
composed of several scaled down and rotated copies of itself. would adapt itself on different hardware architectures. For
Random fractals include that fractal which has an additional instance this architecture would provide a very efficient, robust
element of randomness, allowing for simulation of natural and fast broad-phase that would work on several GPUs. Then
phenomenon. So they exhibit the property of statistical self architecture would provide a narrow-phase working on multi-
similarity. There are there common techniques for generating cores.
fractals are: We have designed a novel view of a tri-dimensional collision
Escape-time fractals: These are defined by a recurrence relation detection pipeline (cf. Figure 8). The sequential pipeline has
at each point in a space (such as the complex plane). Examples been revisited as a parallel pipeline working with broad and
of this type are the Mandelbrot set, Julia set, the Burning ship exact phase using buffers. Contrary to sequential pipeline, we
fractal and the Lyapunov fractal. propose to add a third dimension. This new dimension is the
Iterated Function Systems: these have a fixed geometric architecture showing that one phase can be done on GPUs and
replacement rule. Cantor set, sierpinski carpet, Sierpinski gaset, another on CPUs or one part of phase. This 3D pipeline might be
peano curve, Koch snowflake,Harter-Heighway dragon curve, T- dimensioned for a cluster or grid architecture. We also imagine
square, menger sponge, are some examples of such fractals[6,7]. running a collision detection system on grid, taking into account
Random fractals: Generated by stochastic rather than latency time. On the overall distributed machines predictive
deterministic processes, for example, trajectories of the algorithms would compute possibly colliding sets of objects and
Brownian motion, levy flight, fractal landscapes and the local machines would compute exact collision.
Brownian tree. The latter yields so-called mass- or dendrites
fractals, for example, diffusion –limited aggregation or reaction- 7. Conclusion
limited aggregation clusters[8].
We have presented in this survey new trends in collision
detection performance quest. With the wide range of available
architecture from GPU to grid, we have shown that developing
new models taking into account computer power is essential to
expect real-time interaction in large-scale environment. The link-
up between virtual reality and computer performance has to be
more and more studied and reinforced. Future algorithms
providing a real-time interaction have to be performed through
the use of GPGPU, multithreads, multi or many-cores processors
with memory and cache handling and cluster or grid architecture.
We also come in the conclusion that the use of the fractal
collision detection algorithm reduces the detection process by
examining on the comparatively smaller part. Still the Work is in
progress later on more fruitful result will surely be coming.
Figure 6: Example of Cauliflower Fractal Showing the Self similarity
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 7
www.IJCSI.org

[16]D.Rochon, A generalized Mandelbrot set for bi-complex

numbers, Fracatals 8(4), 355-368.
[17]D. d'Aulignac K. Sundaraj and E. Mazer. A new algorithm
for computing minimum distance, October 13 2000.
[18] David E. Johnson and Elaine Cohen. A framework for
efficient minimum distance computations. In ICRA, pages 3678–
3684, 1998.
[19] DukSu Kim, Jea-Pil Heo, and Sung eui Yoon. Pccd: Parallel
continuous collision detection. Technical report, Dept. of CS,
KAIST, 2008
[20] Dave Knott and Dinesh K. Pai. Cinder: Collision and
interference detection in real-time using graphics hardware. In
Graphics Interface, pages 73–80, 2003
Figure 7: Example of a new tri-dimensional collision detection pipeline. [21] Daniel S. Coming and Oliver G. Staadt. Velocity- aligned
discrete oriented polytopes for dynamic collision detection. IEEE
6. References Trans. Vis. Comput. Graph, 14(1):1–12, 2008.
[1] Arthur D. Gregory, Ajith Mascarenhas, Stephen A. Ehmann, [22] Dobkin, Hershberger, Kirkpatrick, and Suri. Computing the
Ming C. Lin, and Dinesh Manocha. Six degree- of-freedom intersection-depth of polyhedra.ALGRTHMICA: Algorithmica,
haptic display of polygonal models. In IEEE Visualization, pages 9, 1993
139–146, 2000 [23]Everton Hermann, Francois Faure, and Bruno Raffin. Ray-
[2] Avneesh Sud, Miguel A. Otaduy, and Dinesh Manocha. Difi: traced collision detection for deformable bodies. In GRAPP,
Fast 3D distance field computation using graphics hardware. pages 293–299, 2008.
Comput. Graph. Forum, 23(3):557–566, 2004 [24]E. G. Gilbert, D. W. Johnson, and S. S. Keerthi. A fast
[3] Brian Mirtich. V-clip: Fast and robust polyhedral collision procedure for computing the distance between complex objects
detection. ACM Trans. Graph, 17(3):177–208, 1998. in three-dimensional space. IEEE Journal of Robotics and
[4] Bruno Heidelberger, Matthias Teschner, and Markus H. Automation, 4:193–203, 1988
Gross. Detection of collisions and self-collisions using image- [25] Elmer G. Gilbert and Chong Jin Ong. New distances for the
space techniques. In WSCG, pages 145–152, 2004. separation and penetration of objects. In ICRA, pages 579–586,
[5] Bruno Raffin and Luciano Soares. Pc clusters for virtual 1994
reality. IEEE-VR, pages 215–222, 25-29 March 2006. [26] George Baciu and Wingo Sai-Keung Wong. Image-based
[6]Beno^t B. Mandelbrot,Multifractals and 1/f noise , springer- collision detection for deformable cloth models. IEEE Trans.
Verlag, Berlin,1988. Vis. Comput. Graph, 10(6):649–663, 2004.
[27] Gino Van Den Bergen. Efficient collision detection of complex
[7] Beno^t B. Mandelbrot,Multifractals , discussion paper ;
deformable models using aabb trees. J. Graph. Tools, 2(4):1–13,
fractals , attractors and fractals dimentions, In:O, Gurel 1997.
&O.E.Rossler (Eds.), Bifurcation theory and application in [28] Gino Van Den Bergen. A fast and robust guick
scientific disciplines, Paper presented at aconfrence held in New implementation for collision detection of convex objects. J.
York Academy of sciences 316, (pp.463-464), New york : New Graph. Tools, 4(2):7–25, 1999.
york academy of sciences,1979. [29] G. Vanecek, Jr. Back-face culling applied to collision
[8] Beno^t B. Mandelbrot, self –affine Fractals and Fractals detection of polyhedra. The Journal of Visualization and
Dimension, physica scripta 32 (1985), 257-260. Computer Animation, 5(1), January– March 1994.
[9] Beno^t B. Mandelbrot,Multifractals Fractals : Form, chance [30] Gabriel Zachmann. Rapid collision detection by
and dimensions, freeman,1977. dynamically aligned DOP-trees. pages 90–97, March 1998.
[10] Christer Ericson. Real-time Collision Detection. Morgan [31] Gabriel Zachmann. Optimizing the collision detection
Kaufmann, 2005 . pipeline. In Proc. of the First International Game Technology
[11] C .Mc Mullen, The classification of conformal dynamical Conference (GTEC), January 2001.
systems,Current developments in mathematics, Int’Press [32] H.-O.Petigen and D.Saupe(Eds), The Science of Fractals
Cambridge,MA(1995),323-360.. Images, Springer-Verlag,1988.
[12] ] C .Mc Mullen ,Automorphism of rational maps , [33] H.-O.Petigen and P.H Richter, The beauty of Fractals ,
Holomorphic Function and Moduli Vol I, Math. Sci, Res. Springer-Verlag,1986.
Inst.Publ.10 Springer, New York (1988). [34] H.-O.Petigen and H . Jurgens and D.Saupe(Eds), Chaos and
[13] C. Peterson and G.Ryd, Convergence of rational rays in Fractals, Springer-Verlag,,, New York, Inc., 1992.
parameter spaces In: the Mandelbrot Set : Theme and variations , [35] H.-O.Petigen, H . Jurgens and D.Saupe(Eds), The Language
London mathematical society , Lecture note series 274, of Fractals , Scientific American (August 1990),40-47.
Cambridge University Press(2000),161-172 [36] Jeremie Allard, Valerie Gouranton, Loick Lecointre,
[14] C.A. Pickover , Fractals horizons: The future use of fractals Sebastien Limet, Bruno Raffin, and Sophie Robert. FlowVR: A
, New York : St.Martin‘s press, 1996. middleware for large scale virtual reality applications, August-
[15]Domonic E. reeve, Mandelbrot, Julia sets and non linear September 2004.
mapping in : fractals and chaos (Edited by A.J. crilly, R.A. [37] J. W. Boyse. Interference detection among solids and
earnshaw and H.Jones ,) springer- Verlag, Newyork, Inc., surfaces. Communications of the ACM, 22(1):3–9, January 1979.
1991,35-42. [38] J. Canny. Collision detection for moving polyhedra. IEEE
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 8
www.IJCSI.org

Transactions on Pattern Analysis and Machine Intelligence, [55] M. Rani , Fractals in vedic heritage and fractals carpets ,
8:200–209, 1986. Proceeding of national seminar on history , Heritage and
[39] Jonathan D. Cohen, Ming C. Lin, Dinesh Manocha, and development of Mathematical sciences, Oct.2003.18-20.
Madhav K. Ponamgi. I-collide: An interactive and exact collision [56] M. Rani and M.Kumar ,Superior Julia set ,J.Korean Soc.
detection system for large-scale environments. In SI3D, pages Math.educ.Ser.d8(4)(2004),261-277.
189–196, 218, 1995 [57] M. Rani and V.Kumar ,Superior Mandelbrot set ,J.Korean
[40] J.Milnar, Geometry and Dynamics of quadratic rational Soc. Math.educ.Ser.d8(4)(2004),279-291.
maps, Exper.Math.2(1993),37-83. [58] M. Rani and V.Kumar ,A new experiment with the logistic
[41] J.Milnar, Dynamix in one Complex Variable, Vieweg,1999. function , J . Indian Acad.Math .27(2005). No.1.
[42] ] J.Milnar and tan Lei , ”Sierpinski Carpet” as Julia Set . [59]M.Rees, Positve measure sets of ergodic rational maps,
Apendix f in: Geometry and Dynamics of Quadratic Rational Ann.Sci.Ec.Norm.Sup.19(1986), 383-407.
Maps, Experiments. Math. 2(1993), 37-83. [60]M.Shishikura, Hausdroff dimension of the boundry of the
[43] John D. Owens, David Luebke, Naga Govindaraju, Mark Mandelbrot set and Julia sets , Ann. Math.(2) 147(2), 225-267.
Harris, Jens Krüger, Aaron E. Lefohn, and Timothy J. Purcell. A [61]M.A.Krasnosel’skii, Two observations abut the method of
survey of general-purpose computation on graphics hardware. successive approximations, Uspechi Mat.Nauk 10 (1995), 123-
2007. Warning: the year was guessed out of the URL. 127.
[44]Jon Louis Bentley and Jerome H. Friedman. Data structures [62]Michal L. Lapidus and Michiel Frankenhugsen(Editors),
for range searching. ACMCS, 11(4):397–409, 1979. Fractals Geometry and applications : A Jublie of Benoit
[45]Kimberly a.Roth, Julia sets that are full of Holes , Mandelbrot, Part I, San Diego,California,2002 and Ecole
preprint,2002. Normale Superieure de Lyon, 2001.
[46] Larry Seiler, Doug Carmean, Eric Sprangle, Tom Forsyth, [63]Naga K. Govindaraju, Ming C. Lin, and Dinesh Manocha.
Michael Abrash, Pradeep Dubey, Stephen Junkins, Adam Lake, Fast and reliable collision detection using graphics processors. In
Jeremy Sugerman, Robert Cavin, Roger Espasa, Ed Grochowski, COMPGEOM: Annual ACM Symposium on Computational
Toni Juan, and Pat Hanrahan. Larrabee: a many-core x86 Geometry, 2005.
architecture for visual computing. ACM SIGGRAPH'08 [64] Naga K. Govindaraju, Stephane Redon, Ming C. Lin, and
Transactions on Graphics, 27(3), August 2008. Dinesh Manocha. Cullide: Interactive collision detection
[47]Mark Lewis and Berna L. Massingill. Multithreaded between complex models in large environments using graphics
collision detection in java. In Hamid R. Arabnia, editor, hardware. In M. Doggett, W. Heidrich, W. Mark,andA.
Proceedings of the International Conference on Parallel and Schilling,editors, SIGGRAPH/Eurographics Workshop on
Distributed Processing Techniques and Applications & Graphics Hardware, pages 025–032, San Diego, California,
Conference on Real-Time Computing Systems and Applications 2003. Eurographics Association
(PDPTA'06), volume 1, pages 583–592, Las Vegas, Nevada, [65]P. M. Hubbard. Collision detection for interactive
USA, June 2006. CSREA Press. graphicsapplications. EEETransactionson Visualization and
[48]Min Tang, Dinesh Manocha, and Ruofeng Tong. Multi-core Computer Graphics, 1(3):218–230, September 1995. ISSN 1077-
collision detection between deformable models. In Computers & 2626.
Graphics, 2008. [64] Pablo Jiménez, Federico Thomas, and Carme Torras. 3d
[49]M. Teschner,B. Heidelberger,M. Müller, D. Pomeranets, and collision detection: a survey. Computers & Graphics,
M. Gross. Optimized spatial hashing for collision detection of 25(2):269–285, 2001
deformable objects. In T. Ertl, B. Girod, G. Greiner, H. [65]Overmars. Point location in fat subdivisions. IPL:
Niemann, H.-P. Seidel, E. Steinbach, and R. Westermann, Information Processing Letters, 44, 1992.
editors, Proceedings of the Conference on Vision, Modeling and [66] Przemyslaw Prusinkiewicz, modeling and visualization of
Visualization 2003 (VMV-03), pages 47–54, Berlin, November biological structures, Proceedings of Graphics Interface ’93,
19–21 2003. Aka GmbH. Held in Toronto, Ontario, May1993, 128-137.
[50] M. Teschner, S. Kimmerle, B. Heidelberger, G. Zachmann, [67]Przemyslaw Prusinkieicz and Mark Hammel , Language-
L. Raghupathi, A. Fuhrmann, M.-P. Cani, F. Faure, N. restricted iterated function system , Koch constructions, and L-
Magnenat-Thalmann, W. Strasser, and P. Volino. Collision systems , New Directions for Fractals Modeling in Computer
detection for deformable objects. pages 119–140, September Graphics,Siggraph’94 Course Notes , ACM Press, 1994.
2004. [68] S. Kockara, T. Halic, K. Iqbal, C. Bayrak, and Richard
[51]M. C. Lin and S. Gottschalk. Collision detection between Rowe. Collision detection: A survey. Systems, Man and
geometric models: a survey. In Robert Cripps, editor, Cybernetics, 2007. ISIC. IEEE International Conference on,
Proceedings of the 8th IMA Conference on the Mathematics of pages 4046–4051, Oct. 2007.
Surfaces (IMA-98), volume VIII of Mathematics of Surfaces, [69] Sean Quinlan. Efficient distance computation between non-
pages 37–56, Winchester, UK, September 1998. Information convex objects. In ICRA, pages 3324–3329, 1994.
Geometers. [70] Stéphane Redon, Abderrahmane Kheddary, and Sabine
[52] Ming C. Lin and John F. Canny. A fast algorithm for Coquillart. Fast continuous collision detection between rigid
incremental distance calculation. Technical report, UNC. March bodies. Computer Graphics Forum, 21(3):279–288, September
19 1991. 2002.
[53] Moller. A fast triangle-triangle intersection test. JGTOOLS: [71] Srikanth Bandi and Daniel Thalmann. An adaptive spatial
Journal of Graphics Tools, 2, 1997. subdivision of the object space for fast collision detection of
[54]M. Rani , Iterative Procedure in Fractals and Chaos , Ph.D. animated rigid bodies. Comput. Graph. Forum, 14(3):259–270,
Thesis, Gurukul Kangri Vishwavidalaya, Haridwar.2002. 1995.
IJCSI International Journal of Computer Science Issues, Vol. 7, Issue 1, 2010 9
www.IJCSI.org

[72] Stephen Cameron. Enhancing GJK: Computing minimum

and penetration distances between convex polyhedra, January 27
1997.
[73] Susan Fisher and Ming C. Lin. Fast penetration depth
estimation for elastic bodies using deformed distance fields,
UNC. July 17 2001.
[74] Stefan Gottschalk, Ming Lin, and Dinesh Manocha.
Obbtree: A hierarchical structure for rapid interference detection.
In Proceedings of the ACM Conference on Computer Graphics,
pages 171–180, New York, August 4–9 1996. ACM.
[75] S.Morosawa, Y. Nishimura, m. Tangiguchi and T. Ueda,
Holomorphic Dynamics, Cambridge University Press(2000).
[76] Sri Pinggalacarya, chandos Sastrum (Commentry by Sri
Halyadha Bhatta and edited by pandita Kedarnatha,
Chaukhambha publishers, varansi,2002)
[77] Tan lei , Similarity between the Mandelbrot sets and Julia
sets, Commun. Maths.phys.134(1990),587-617.
[78] V.a. Norton , Julia sets in quaternions, Computer and
Graphics 13,2(1989)267-278.
[79]W.r.Mann , Mean value methods in iteration , Proc.
Amer.math. Soc. 4(1953),506-510.
[80] Yu.A. Shashkin , Fixed points, Amer.Math.Soc,1991.
[81] Young J. Kim, Ming C. Lin, and Dinesh Manocha. Deep:
Dual-space expansion for estimating penetration depth between
convex polytopes. In ICRA, pages 921– 926. IEEE, 2002.
[82] Y. Sato, M. Hirata, T. Maruyama, and Y. Arita. Efficient
collision detection using fast distance- calculation algorithms for
convex and non-convex objects. In Proc. IEEE Intern. Conf. on
Robotics and Automation (Minneapolis, Minnesota 1996), pages
771–778. IEEE, 1996.