Computing the Minimum Directed Distances Between Convex Polyhedra
|
|
- Nathaniel Butler
- 6 years ago
- Views:
Transcription
1 JOURNAL OF INFORMAION COMUING SCIENCE DISANCES AND ENGINEERING BEWEEN CONVEX 15, OLYHEDRA (1999) 353 Computng the Mnmum Drected Dstances Between Convex olyhedra Department of Electrcal Engneerng Natonal awan Insttute of echnology ape, awan 106, R.O.C. Gven two dsjonted objects, the mnmum dstance (MD) s the short Eucldean dstance between them. When the two objects ntersect, the MD between them s zero. he mnmum drected Eucldean dstance (MDED) between two objects s the shortest relatve translated Eucldean dstance that results n the objects comng just nto contact. he MDED s also defned for ntersectng objects, and t returns a measure of penetraton. Gven two dsjonted objects, we also defne the mnmum drected L dstance (MDLD) between them to be the shortest sze ether object needs to grow proportonally that results n the objects comng nto contact. he MDLD s equvalent to the MDED for two ntersectng objects. he computaton of MDLD and MDED can be recast as a Mnkowsk sum of two objects and fnshed n one routne. he algorthms developed here can be used for collson detecton, computaton of the dstance between two polyhedra n three-dmensonal space, and robotcs path-plannng problems. Keywords: mnmum dstance, mnmum drected dstance, Mnkowsk sum, collson detecton, path plannng 1. INRODUCION Determnng the mnmum dstance between two convex polyhedra s an mportant problem n robotcs, mage processng, CAD systems, computatonal geometry, and other areas of nformaton processng whch deal wth geometrcal data. All work that s developed for automated path plannng requres at ts lowest level the ablty to detect whether or not collson has occurred. he ablty to compute dstance and nterference effcently wll result n a substantal reducton n the overall tme requred for most path-plannng algorthms. In general, a path plannng algorthm needs to ascertan for any poston n the workspace not only f a collson has occurred, but also how close t s to occurrng f t has not. For nstance, a generate-and-test type path-plannng algorthm requres tests for whether confguraton and trajectory are collson-free, and methods for searchng ntermedate subgoals [2]. Several technques have been reported for calculaton of the mnmum dstance between convex polyhedra. In partcular, Glbert, et al. [5] defned an object by means of the convex hull of ts vertces and provded a quck procedure to compute the dstance between 3D objects. her approach performs an teratve sequence of dstance mnmzatons to obtan elementary subsets of the orgnal shape untl a subset contanng the global mnmum s attaned. Bobrow [1] proposed an approach whch casts the problem as a constraned Receved January 4, 1997; accepted September 5, Communcated by Youn-Long Ln. 353
2 354 nonlnear mnmzaton. Hs algorthm uses a drect approach to mnmze the nonlnear dstance functon, whch generates a sequence of search drectons along the surfaces of the objects to obtan the global mnmum. he computaton tme of ths algorthm s roughly lnear wth the number of faces. Ln and Canny [6] proposed an ncremental dstance calculaton startng wth a canddate par of features, one from each polyhedron, that checks the closest ponts that le n these features. he algorthm then steps to the next closest par of features untl the closest ponts are found. he mnmum dstance between three-dmensonal segments was developed n [8]. he computatonal effcency of two-dmensonal algorthms for polygons was gven n [4,10]. he dstance between boxes was consdered n [9]. he above mnmum dstance algorthms are asymptotcally fast, but they only return a zero value for ntersectng objects. Dstance s used as a measure of how far a robot part s from colldng wth an obstacle. When the two objects ntersect, the dstance between them s zero. hs gves no nformaton about the ntensty of the ntersecton. he general objectve of a penetraton measure s to quantty the depth of ntersecton for object modelng. A mnmum drected Eucldean (or translated) dstance was proposed n [2,3] to defne the ntensty of penetraton and uses the negatve of the mnmum Eucldean dstance by whch the two overlappng objects must be relatvely translated so as to have no nteror pont n common. he mnmum drected Eucldean dstance s equvalent to the dstance between the two objects f the objects are dsjonted. Buckley [2] proposed a method to compute the mnmum drected Eucldean dstance between two-dmensonal convex polygons and used ths measure n a penalty approach n collson-avodance robot moton plannng. he result shows that the nonntersecton constrant provdes useful nformaton n the case of body ntersecton, and that path plannng can be done accordng to the flexble trajectory paradgm. Cameron & Culley [3] appled the Mnkowsk sum technque to compute the mnmum translated dstance between two convex polyhedra n a three-dmensonal space. However, the complexty and completeness of ther algorthm was not analyzed. As one step toward reducng the number of computatons needed for path plannng, the am of ths research s to present effcent algorthms to compute the mnmum drected dstance functons between two convex polyhedra. he polyhedral representaton of 3D objects s wdely used n robotcs and computer graphcs research. We defne a new mnmum drected L dstance (MDLD) between convex polyhedra and derve ts relatonshp wth the mnmum drected Eucldean dstance (MDED) proposed n [2,3]. In ts smple form, the measure of the L dstance between a pont and a convex polyhedron s the maxmum of the dstances of the pont from the half-space whch passes through the faces of the polyhedron. Our computaton of the MDLD and MDED functons and the Mnkowsk sum of two convex polyhedra are based on the boundary model of the polyhedron. he followng s an outlne of the paper and ts contents. In Secton 2, the mnmum drected L dstance s defned, and ts propertes as well as ts connecton wth the mnmum drected Eucldean dstance are also derved. Effcent algorthms for computng both MDLD and MDED are proposed n Secton 3. Several examples and applcatons are llustrated n Secton 4. Fnally, Secton 5 provdes conclusons wth regard to the proposed approach.
3 COMUING DISANCES BEWEEN CONVEX OLYHEDRA ROBLEM FORMULAION AND RELIMINARIES We wll assume that two objects and A are convex polyhedra n R r space for further dscusson. When r = 2, objects and A are two-dmensonal convex polygons, and when r = 3, objects and A are three-dmensonal convex polyhedra. Some termnology wth regard to polyhedral objects used n ths paper are ntroduced here. A convex polyhedron n three-dmensonal space s characterzed by ts faces, edges, and vertces. A face of a threedmensonal polyhedron s a 3D convex polygon. A plane n 3D ndcates a set whch satsfes the plane equaton, n x+d = 0. A lne means a straght lne of nfnte length passng through two ponts a and b, and lne segment ab ndcates a segment of a lne connectng two end ponts a and b. An edge s a lne segment connectng two vertces. For twodmensonal polygons, the terms face and edge are usually nterchangeable. he closest par of features between two objects s defned as the par of features whch contan the par of closest ponts between objects. he par of closest ponts means the mnmum Eucldean dstance ponts between two objects. Let p (p Œ R r ) be the translatonal vector of the reference pont of polyhedron. A convex polyhedron wth m faces can be represented by the set where = {x Œ R r ÍN (x p) + d 0}, N = [n 1,..., n m ] Œ R r m, and d = [d 1,..., d m ] Œ R m. Matrx N s an r by m matrx whose th column n s the unt outward normal vector of the th face of polyhedron. Vector d s an m dmensonal vector, and d s the magntude of the vector from reference pont p perpendcular to the th face, measured n the negatve n drecton. Let the vertces of be represented by U = [u 1,..., u k ] Œ R r k, where k s the number of vertces, and u 1,..., u k are the vertex translatonal vectors relatve to the reference pont of. A grown convex polyhedron of convex polyhedron s defned as = {x Œ Rr ÍN (x p) + d s, s Œ R m and s 0}. and a proportonally grown polyhedron e (e Œ R, e 0) of convex polyhedron s defned by e = {x Œ R r ÍN (x p) + d [1,..., 1] e}. Smlarly, a convex polyhedron A wth n faces and l vertces s represented by the set where A = {x Œ R r ÍA (x q) + b 0},
4 356 A = [a 1,..., a n ] Œ R r n, b = [b 1,..., b n ] Œ R n. Moreover, q (q Œ R r ) s the translatonal vector of the reference pont of A, and ts vertces are represented by V = [v 1,..., v l ] Œ R r l, where v 1,..., v l are the vertex translatonal vectors of A relatve to ts reference pont. he followng defntons and notatons are used n the paper and are defned below. Defnton 1: he maxmum value of vector y, max(y), s gven by max(y) = max{y 1,..., y n }, y Œ R n. Defnton 2: he nteger functon arg{max(y)} s the ndex (or argument) such that y = max(y). Defnton 3: he mnmum value of vector y, mn(y), s gven by mn(y) = mn{y 1,..., y n }, y Œ R n. Defnton 4: he nteger functon arg{mn(y)} s the ndex (or argument) such that y = mn(y). -: polygon whose vertex vectors are the negatons of the vertex vectors of when the reference pont of s at the orgn. A : Mnkowsk sum of and -A, and t s the set of translatons of the reference pont q of Q that brng t nto nterference wth and has ts new reference pont at pont p,.e. A = {q : «A π }. d(x, ): mnmum dstance (MD) between pont object x and object, and dx (, )= mn x y 2 y d (x, ): mnmum drected L dstance (MDLD) between a pont object x and an object, and d (x, ) = max(n (x - p) + d) = -mn(-n (x - p) - d). A contour of d (x, ) = e gves all the boundary faces of e for e greater than zero. When x œ, MDLD d (x, ) (= max(n (x p) + d ) > 0) s the shortest sze that needs to grow proportonally, resultng n pont x beng on the boundary of e, and that resembles an L - norm dstance. When x Œ, d (x, ) (= -max(-n (x - p) - d ) 0) s the negatve of the shortest dstance that x needs to translate, resultng n pont x just touchng the boundary of. Note that d (x, ) Á x=q = max(n (q - p) + d ) = max(-n (p - q) + d ) = d (x, c )Á x=p, where c = {x Œ R r Í-N (x - q) + d 0} s the Mnkowsk sum of pont object q and object. d 2 (x, ) : mnmum drected Eucldean dstance (MDED) between pont x and object, and d (, d ) ( x, ) x 2 x = d ( x, ) x. {
5 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 357 Examples of MD, MDLD and MDED between pont object x and convex polygon n two-dmensonal space are shown n Fg. 1. MDLD d (x, ) returns the orthogonal dstance from pont x to the half-space determned by arg{max(n (x - p) + d )}, and d 2 (x, ) returns the mnmum dstance from pont x to the correspondng closest face. Assume that the orthogonal projecton of pont x to the closest plane s denoted by x *. If pont x * s on the nearest face, then d 2 (x, ) = d (x, ); otherwse, d 2 (x, ) > d (x, ). hus, n general, we have the relatonshp d 2 (x, ) d (x, ). In other words, d (x, ) determnes only the dstance between the closest par of features of a pont and a face. However, d 2 (x, ) determnes the dstance between the closest par of features n all cases, ncludng a par of a pont and an edge, and a par of a pont and a vertex. MDED = MD x MDED = MDLD < 0 MDLD MD = 0 x (a) x (b) x Fg. 1. Geometrcal llustraton of MD, MDLD, and MDED between a pont object x and a convex polygon n two-dmensonal space; (a) pont x s outsde of, and (b) pont x s nsde of. We can now defne the mnmum drected dstances (MDD) between two convex polyhedra A and. In concept, the dstance problem for two objects A and can be reduced to the problem of fndng the dstance from a pont object and the Mnskows sum A. d(a, ): mnmum dstance (MD) between convex polyhedra A and, and da (, ) = mn x y = d( x, A ) =. 2 x q x A, y d (A, ): he mnmum drected L dstance (MDLD) between convex polyhedra A and, d (A, ), s gven by d (A, ) = d (x, A )Á x = q. Note that d (A, ) = d (x, A )Á x = q = d (x, A )Á x = p = d (, A), (1) where A s the Mnkowsk sum of A and -. hus, d (A, ) s a symmetrc dstance functon. Fg. 2 shows an example of Eq. (1) n 2D. When A and are dsjonted, d (A, ) s the shortest sze ether object needs to grow proportonally that results n the objects beng n contact and, hence, returns a measure of dstance between A and. In the two-dmensonal case, d (A, ) can be geometrcally nterpreted as
6 358 q d ( A, ) A A p (a) d (a) d ( A,) = d ( q, A ) (A, ) = d (q, A ) A q A d (, A) p (b) d (b) d (, A) = d ( p, A p ) (, A) = d (p, A p ) Fg. 2. Illustraton of d (A, ) = d (x, A )Á x = q = d (x, A )Á x = p = d (, A), n 2D; (a) d (A, ) = d (q, A ), and (b) d (, A) = d (p, A ). d (A, ) = max{e A, e }, for A «=, (2) where e A s the smallest value of e such that A e and ntersect, and e s the smallest value of e such that A and e ntersect. If the objects overlap, d (A, ) s the negatve of the shortest dstance needed to translate one object wth respect to the other untl there s no ntersecton. Fg. 3 shows the geometrcal nterpretaton of MDLD n 2D. d 2 (A, ): mnmum drected Eucldean dstance (MDED) between convex polyhedra A and, and da (, ) fa d2( A, ) = =. d ( A, ) f A Examples of MDED are also shown n Fg. 3. One computatonal dffculty n the above formulaton s that the Mnkowsk sum generally has a very complcated shape n hgher dmensons. Nevertheless, not every face of the set A needs be consdered to fnd the mnmum dstance. hs s the focus of the followng secton.
7 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 359 A ε A ε A ε MD ε (a) (a) and and A A and are dsjonted A MD = 0 MDED=MDLD<0 (b) and A are overlappng (b) and A are overlappng Fg. 3. Geometrcal nterpretaton of MDLD, MD, and MDED between two convex polygons A and n 2D; (a) dsjonted case, and (b) overlappng case. 3. MDD ALGORIHMS In ths secton, we shall develop effcent algorthms for computng the mnmum drected dstance functons for two convex objects. We shall frst develop the boundarymodel approach to represent the Mnkowsk sum A, whch s most sutable for computng the proposed mnmum drected dstances between convex polyhedra. If A and are convex polyhedra, then set A s also a convex polyhedron, wth ponts on ts boundary representng the confguraton wth whch objects A and come nto contact. he faces of A correspond to modes of contact between the objects such that they can slde along each other, and these modes can be classfed as (1) face-vertex (FV) contact, (2) vertex-face (VF) contact, and (3) edge-edge (EE) contact. hs set of three types of nteracton cover all possble types of contact between polyhedra. Other forms of contact are specal cases. For nstance, an edge-face contact requres at least two contacts from among the above three contact types. For FV contact cases, a supportng face of A s obtaned when a face of sldes aganst a vertex of A. Let
8 360 n ( x p) + 0 (3) d be the half-space correspondng to a face of ; then, the vertex of A wth a maxmum value of ( n vj ) for j = 1,..,l, can slde aganst (see Fg. 4(a)), where l s the number of vertces of A. hus, the set A should satsfy n ( x p) + d e, = 0, 1,..., m 1, (4) where e = max{ nv, j = 1,..., l} = mn( V n ), j and m s the number of faces of. For each th face of, let j = arg(mn(v n )); the th face of and jth vertex of A are called a face-vertex (FV) contact par, and the correspondng mnmum dstance d between them s δ = n ( q p) + d + mn( V n ) = n ( q p) + d e. (5) he contact par whch has a maxmum value of d n Eq. (5) among all possble face-vertex contact pars s called a face-vertex closed par. For VF contact cases, a supportng face of A s obtaned when a face of A sldes aganst a vertex of (see Fg. 4(b)). Let n 1 ( x-p ) + d 1= e 1 n 1 ( x-p ) + d 1= 0 n 1 n 1 p n 1 h 1 q A h h 3 2 e 1 e 1 (a) FV contact -n 1 h2 = = max{ -n 1 h 1 -n 1 h2 -n 1 h 3 } u 1 u 2 p u 3 f 2 a 2 -b 2 a 2 ( x-q ) + = 0 A q b 2 -a 2 ( x-p ) + b 2 = f 2 f 2 -a 2 a 2 = - u 3 = max{ - u 1 - u 2 - u 3 } a 2 (b) VF contact Fg. 4. Illustraton of the supportng faces of Mnkowsk sum A generated by (a) face-vertex contact, and (b) vertex-face contact. a 2 a 2
9 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 361 a ( x q) + 0 b (6) correspond to a face of A; then, the vertex of wth a maxmum value of ( a uj ) for j = 1,.., k, can slde aganst A, where k s the number of vertces of. hus, A should also satsfy where a ( x p) f b, = 0, 1,..., n 1, (7) f = max{ a u, j = 1,..., k} = mn( U a ), j and n s the number of faces of A. For each th face of A, let j = arg(mn(u a )); the th face of A and jth vertex of are called a vertex-face (VF) contact par, and the correspondng mnmum dstance d between then s δ = a ( p q) + b + mn ( U a ) = a ( p q) + b f. (8) he contact par whch has a maxmum value of d n Eq. (8) among all possble vertex-face contact pars s called a face-vertex closed par. he three-dmensonal case s exactly analogous for the FV and VF contacts n the two-dmensonal case. he EE contact mode occurs only n the three-dmensonal case. In EE contact cases, not every par of edges wll generate a supportng face of A. Let u a u b be an edge of, and let v g v l be an edge of A; then, edges u a u b and v g v l can come nto contact f there exsts a normal vector n (see Fg. 5) n = uu uu α β γ λ uu α β uu (9) γ λ 2 n n = e A x e p A q e A e 2 e p n. ( x-p ) = e 1 + e 2 e 1 p Fg. 5. Illustraton of the supportng faces of the Mnkowsk sum A generated by edge-edge contact n 3D.
10 362 such that and e 1 = n u a = n u b = max{n u j, j = 1,..., k} 0 e 2 = -n v g = -n v l = max{-n v, = 1,..., l} 0. In addton, the supportng face generated by the EE contact s the half-space represented by n (x - p) e 1 + e 2. (10) hese two edges become an edge-edge contact par, and the mnmum dstance d between the two lnes u a u b and v g v l s d = n (q - p) - e 1 - e 2. (11) he edge-edge par whch has a maxmum value of d among all possble edge-edge contact pars s called an edge-edge closed par (named edge of and edge j of A). For two-dmensonal convex polygonal objects, we only need to consder the FV contact for all faces of and the VF contact for all faces of A; thus, an analytcal form of the set A can be obtaned. For cases of FV contact, we obtan the grown object : = {xœr r ÍN (x - p) + d - e 0}, where m e = max{ nv, j = 1,..., l} = max( Vn), e R. j For the VF contact case, we obtan the grown object A c : where A c = {xœr r Í-A (x - p) + b - f 0}, n f = max{ au, j = 1,..., k} = max( Ua), f R. j he set A s, thus, the ntersecton of and A c and s the set r A = R + x N A ( x p) d e b f 0. (12) In addton, the set A s, then, the set represented by A p r = R + x N A ( d e x q ) b f 0. (13)
11 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 363 Fg. 6 shows an example of the two-dmensonal Mnkowsk sum A obtaned by the boundary-model approach. For the three-dmensonal case, the expresson n (12) s only a good approxmaton for A. a 2 h 1 h2 h 3 q A a 1 n 2 u 1 u 2 p n 1 u 3 a 3 n 3 (a) (a) Object Object and and AA -a 3 u1 -n 2 h 3 -a1 u2 n 1 -A q p n - h n 1 h2 -a2 u3 A c (b) (b) Mnkowsk Mnkowsk sum sum A A Fg. 6. An example of the two-dmensonal Mnkowsk sum A ; (a) the objects A and, and (b) A obtaned by the boundary-model approach. For two dsjonted convex polygons n 2D, there are three possble types of closest pars of features: (1) a par of vertces, (2) a vertex and an edge, and (3) a par of edges. Case (3) occurs when the closest par of features s a par of parallel edges, and the mnmum dstance between them s easy to obtan. Case (2) (a vertex and an edge) s mportant
12 364 because t covers case (1) (a par of vertces). Algorthm MDD2, whch computes the MDLD, MD, and MDED functons for two-dmensonal convex polygons, works n the followng way. Frst, t computes the MDLD functons between pont q and objects for and A c, and searches for the closest par of features from the vertex-face closed par and the face-vertex closed par. If objects A and are overlappng, there s no need to contnue. Otherwse, t fnds the closest par of features from the face-vertex closed par and vertex-face closed par, and then checks whether the closest par of features s a par of vertces or not. Detals of algorthm MDD2 are stated below. Algorthm MDD2: Mnmum drected dstances for two convex polygons A and. Input: convex polygons A and Output: MD d(a, ), MDLD d (A, ), and MDED d 2 (A, ) Intal: x = q ; reduce A to a pont Step 1: Compute d (q, ) ; face-vertex closed par Step 2: Compute d (q, A c ) ; vertex-face closed par Step 3: d (A, ) = max{d (q, ), d (q, A c )} If (d (A, ) 0 ), then ; overlappng case d 2 (A, ) = d (A, ) d(a, ) = 0 otherwse, f (d (q, ) > d (q, A c )), then ; dsjonted case = arg{max(n (q - p) + d - e)} j = arg{mn(v n )} d(a, ) = d(vertex j of A, edge of ) ; face-vertex closed par d 2 (A, ) = d(a, ) or f (d (q, ) < d (q, A c )), then = arg{max(-a (q - p) + b - f)} j = arg{mn(u a )} d(a, ) = d(vertex j of, edge of A) ; vertex-face closed par d 2 (A, ) = d(a, ) or f (d (q, ) = d (q, A c ), then ; dsjonted case = arg{max(n (q - p) + d - e)} j = arg{max(-a (q - p) + b - f)} d(a, ) = d(edge of, edge j of A) ; parallel edges d 2 (A, ) = d(a, ) Assume that an operaton means an nner-product of two vectors n R r space. he frst step n Algorthm MDD2 requres (m+ml) operatons, the second step requres (n+nk) operatons, and the last step requres four operatons to solve the smple problem of comput-
13 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 365 ng the mnmum dstance between a pont and a lne segment. hus, the computatonal complexty of MDD2 s O(ml+nk). he complexty can be further reduced to O(m+n) as wll be dscussed next. he search for m FV contact-pars n Step 1 of MDD2 can be done n lnear tme for convex polygons and A as follows. Assume that the th face of and the jth vertex of A s a FV contact-par; then, t s requred that or nv j nv t, t = 0, 1,..., n 1, n ( vt vj) 0, t = 0, 1,..., n 1, where n s the number of vertces of polygon A. Because of the convexty of polygon, the above nequalty constrants can be reduced to the followng two constrants: and n ( v j vj) 1 0 n ( v j+ vj) 1 0 he procedure of algorthm FVC for searchng m FV contact-pars s gven n detal n the followng. Algorthm FVC: Input: wo convex polygons and A, where and A contan m and n faces, respectvely. Output: m FV contact-pars. Step 1: {Fnd the FV contact-par of n 0.} 1.1 regster FV = for t = 0 to n-1 do 1.3 FV = max{fv, n 0 v t }; 1.4 f FV = n 0 v t then s = t; 1.5 end for Step 2: {Fndng the FV contact-par of n, where = 1, 2,..., m 1.} 2.1 t = s 2.2 for = 1 to m 1 do 2.3 whle n (v t-1 - v t ) < 0 or n (v t+1 - v t ) < 0 do 2.4 t = (t + 1) mod n; 2.5 end whle 2.6 end for he Algorthm FVC for searchng m FV contact-pars conssts of two major steps. In the frst step, we fnd the contact-par of n 0 of by searchng each vertex of A. hus, ths step takes O(n) tme to fnd max{ n 0 v t, t = 0, 1,..., n 1} for the face n 0 and ts correspondng contact-par v s, where s = arg(max{ n 0 v t, t = 0, 1,..., n 1}). In the second step, we utlze the property descrbed above to fnd the contact-par of the faces n 1, n 2,..., n n 1 wth ther correspondng contact-pars smply by sequentally searchng the vertces v s, v s+1,..., and v s n ccw order. hs step has O(m) tme to fnd the contact-pars of faces n 1, n 2,..., n n 1 of. herefore,
14 366 the algorthm FVC for computng the FV contact-pars has O(n + m) tme complexty. For the VF contact case, the n VF contact-pars can be computed n a smlar way. hus, the procedure used to compute the FV and VF contact-pars of two polygons and A has O(n + m) tme complexty. herefore, the computatonal complexty of MDD2 s reduced to O(n + m). For two nonoverlappng convex polyhedra n three-dmensonal space, there are sx possble types of closest pars of features: (1) a par of vertces, (2) a vertex and an edge, (3) a vertex and a face, (4) a par of edges, (5) an edge and a face, and (6) a par of faces. Cases (5) and (6) occur when the closest par of features s two parallel faces and/or edges. Cases (3) and (4) are most mportant for computng the MDLD functon. Cases (1) and (2) are degenerated cases of Cases (3) and (4) and only need to be consdered n computng the MD and MDED functons. he MDD3 algorthm whch computes MDLD, MD, and MDED for two three-dmensonal convex polyhedra A and works n a way smlar to that of algorthm MDD2 except that t consders addtonal edge-edge pars, and t s lsted below. Algorthm MDD3: Mnmum drected dstances for two convex polyhedra A and. Input: convex polyhedra A and Output: MD d(a, ), MDLD d (A, ), and MDED d 2 (A, ) Intal: x = q ; reduce A to a pont Step 1: Compute d (q, ) ; face-vertex closed par Step 2: Compute d (q, A c ) ; vertex-face closed par Step 3: Compute d max among edge-edge pars ; edge-edge closed par Step 4: d (A, ) = max{d (q, ), d (q, A c ), d max } If ( d (A, ) 0 ), then ; overlappng case d 2 (A, ) = d (A, ) d(a, ) = 0 otherwse, f (d (q, ) > d (q, A c ), d max ), then ; dsjonted case = arg{ max( N (q p) + d e)} j = arg{ mn(v n ) } d(a, ) = d(vertex j of A, face of ) ; face-vertex closed par d 2 (A, ) = d(a, ) or f (d (q, A c ) > d (q, ), d max ), then = arg{ max ( A (q p) + b f)} j = arg{ mn(u a ) } d(a, ) = d(vertex j of, face of A) ; vertex-face closed par d 2 (A, ) = d(a, ) or f (d (q, A c ) = d (q, ) > d max ), then = arg{ max (N (q p) + d e)} j = arg{ max( A (q p) + b f)} d(a, ) = d(face of, face j of A) ; parallel faces d 2 (A, ) = d(a, ) or f (d max > d (q, ), d (q, A c )), then d(a, ) = d(edge of, edge j of A) ; edge-edge closed par d 2 (A, ) = d(a, ) or f ( d max = d (q, ) > d (q, A c )), then = arg{ max (N (q p) + d e)}
15 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 367 d(a, ) = d(edge j of A, face of ) d 2 (A, ) = d(a, ) or f (d max = d (q, A c ) > d (q, )), then j = arg{ max ( A (q p) + b f)} d(a, ) = d(face j of A, edge of ) d 2 (A, ) = d(a, ) ; parallel face and edge ; parallel edge and face he above algorthm frst computes the MDLD functon between pont q and set and searches for the vertex-face closed par. Second, t computes the MDLD functon between pont q and set A c and searches for the face-vertex closed par. It then fnds the maxmum value d max of the mnmum dstances n Eq. (11) among possble c edge-edge pars whch are generated durng the frst two steps, as shown n Fg. 7. he edges can become a contact par f each les n the boundary of ts face-vertex closed par and vertex-face closed par found durng Steps 1 and 2. If objects A and are overlappng, then t stops. Otherwse, t fnds the closest par of features among the closed pars found n Steps 1, 2, and 3, and then checks whether the closest par of features s a par of vertces or a par contanng a vertex and an edge. vertex-face closed par face-vertex closed par A Fg. 7. ossble canddate edge-edge pars for computng MDLD. he canddate edge-edge par les on the boundary of ts face-vertex closed par and vertex-face closed par. For the MDD3 algorthm, the frst step requres (m+ml) operatons, the second step requres (n+nk) operatons, and the thrd step requres around c(l+k) operatons. he last step requres many fewer operatons than does Step 3 to compute the mnmum dstance between a pont and a 3D convex polygon or between two lne segments [8]. he problem of fndng the mnmum dstance between two segments can be transformed nto the problem of fndng the mnmum dstance between a pont and a parallelogram n 3D. Computng the mnmum dstance between a pont and a convex polygon n 3D s a smple task and can be solved n at most max (l,k) operatons. hus, the computatonal complexty of MDD3 s O
16 368 (ml+nk+c(l+k)). he complexty can be further reduced to O(m+n+c(l+k)) when both polyhedra A and have fxed orentatons durng moton, and growth obstacles and A c are only computed once. Although the FV constact pars n Algorthm MDD3 cannot be done n lnear tme as we have shown n Algorthm MDD2, the computaton of MDLD usng the boundary model of the Mnkowsk sum only requres consderaton of cases wth FV, VF, and EE contacts. Compared to exhaustve searchng, the number of operatons requred to compute MD and MDED s reduced to about one-thrd. Moreover, the penetraton measure for two ntersectng objects can also be obtaned usng the proposed approach. 4. EXAMLES AND ALICAIONS he examples used n testng algorthm MDD3 nclude platonc solds lsted n able 1. he data structure of a convex polyhedron has a feld for ts faces, edges, vertces, the poston of the reference pont, and ts orentaton. Each face s represented by ts outward normal vector and ts dstance from the orgn. Its data structure also ncludes a lst of vertces whch le on ts boundary. Each edge s descrbed by two connectng vertces and ts two neghborng faces. Each vertex s characterzed by ts x, y, and z coordnates wth respect to ts reference pont. Examples were run on an IBM C whch had a 33Mhz 486 CU. Fg. 8 shows two examples of the mnmum drected Eucldean dstance d 2 (A, ) obtaned usng the MDD3 algorthm. able 1. latonc solds used n 3D examples. etrahedron Cube Octahedron Icosahedron Number of vertces Number of faces (a) cube and cosahedron (b) cube and cube Fg. 8. Example results of algorthm MDD3 for the mnmum dstance between two platonc solds; (a) cube and cosahedron, and (b) cube and cube.
17 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 369 he resultng mnmum dstance has been verfed by usng the tradtonal optmzaton (reduced gradent) routne. he computatonal tme needed for the MDD3 algorthm for objects lsted n able 1 s summarzed n able 2. If sets and A c were computed n advance, the subroutne could generally be run n several mllseconds (about 3 to 8 mllseconds). Numercal data n able 2 ndcate that algorthm MDD3 roughly ncreased lnearly n computaton tme wth the total number of vertces and faces. able 2. Runnng tme (n mllseconds) of MDD3 between two platonc solds A and wth / wthout computng and A c. A \ etrahedron Cube Octahedron Icosahedron etrahedron 7.44 / / / / 6.1 Cube / / / / 7.2 Octahedron 9.78 / / / / 7.0 Icosahedron / / / / 8.4 In the frst applcaton, we wll llustrate the procedure used n applyng the proposed MDLD functon to detect the ntersecton between a lne segment and a convex polygon n a two-dmensonal x-y space. he lne segment may represent a lne path of a pont object whle the polygon represents ether a polygon obstacle or a confguraton-space obstacle. he lne segment n two-dmensonal space can be represented by a degenerated polygon as shown n Fg. 9(a). hus, the ntersecton detecton problem becomes a standard problem whch algorthm MDD2 can solve. Consder a lne segment wth two end ponts p 0 and p 1, and a unt vector u whch ponts to pont p 1 from pont p 0, as shown n Fg. 9(a). Lne segment p 0 p 1 n 2D can be represented by a polygon wth two vertces, p 0 and p 1, and satsfyng two planar half-spaces as below: n = z u x p 1 p 1 u u p 0 -n p 0 (a) (a) 2D case (b) 3D case Fg. 9. Collson detecton between a lne segment and a convex polyhedron ; (a) 2D case, and (b) 3D case.
18 370 n (x - p 0 ) 0 and -n (x - p 0 ) 0, where n = z u, s the vector cross-product operaton, and z s the z-axs unt vector. herefore, the collson ntersecton problem between a lne segment and a convex polygon s reduced to that of fndng the MDLD. hs methodology can be drectly appled to threedmensonal cases, where a lne segment n three-dmensonal space s represented by a convex polyhedron havng one edge, p 0 p 1, and two vertces, p 0 and p 1, as shown n Fg. 9(b). In the second applcaton, we wll llustrate a generate-and-test 2D path-plannng algorthm for a pont robot by usng the proposed MDLD as a key functon. he nput s a set of statonary convex polygonal obstacles, the ntal poston, and the fnal poston. If the robot s not a pont object, then we can reduce the problem to a pont robot problem by workng on the confguraton space. he procedure s as follows. Algorthm 2DAH: 2D path plannng Input: ntal and fnal postons Output: Collson-free lne-segment path passes the subgoals n the subgoal lst {S 0 = ntal,s 1,..., S N = fnal } Intal: startng pont S = ntal, and target pont G = fnal, subgoal lst = {S 0 = ntal } Step 1: Detect ntersecton wth a lne segment path. Connect startng S pont and target G pont. Identfy the obstacles ntersectng the straght lne path between S and G. If no obstacle s found, then go to Step 2; otherwse go to Step 3. Step 2: Update subgoal lst. If target pont G s equal to goal, then place goal nto the subgoal lst and fnsh. Otherwse, place target G nto the subgoal, let the new startng pont S = G and the new target pont G = goal, and go to Step 1. Step 3: Search the nearest obstacle. Search the nearest obstacle among colldng obstacles to the startng pont and go to Step 4. he nearest obstacle s the one whch has the mnmum dstance to startng pont S. Step 4: Group connectng obstacles. Does the nearest obstacle ntersect other obstacles? If so, add these obstacles to the nearest obstacle lst. Repeat addng of those obstacles whch ntersect obstacles n the lst to the nearest obstacle lst. Go to Step 5. Step 5: Select mdpont. Straght lne L, whch passes through S and G, dvdes all the vertces n the nearest obstacle lst nto two groups. Search the farthest vertex from the group whch has the smaller average dstance from vertces to the lne L. he farthest pont s the one wth the largest dstance to straght lne L. Assgn the selected farthest vertex as the new target pont G and go to Step 1.
19 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 371 In the above algorthm, Steps 1, 3, and 4 are easly solved by the MDLD functon developed here. It s noted that Step 4, whch connects overlappng obstacles, can be done n advance and requres very lttle tme durng path plannng. An example of the result of the proposed algorthm n selectng one subgoal s shown n Fg. 10. he 2DAH algorthm s qute fast, and the planned trajectores are near the shortest paths. Fg. 11 shows a result of path plannng done by the 2DAH algorthm wth only seven vertces vsted and ffteen lne segments generated. Algorthm 2DAH was tested n the envronment as shown n Fg. 11 wth 100 randomzed ntal and fnal ponts. he average runnng tme on an IBM C was about 45 mllseconds. he algorthm proposed here dffers from prevous approaches n three mportant ways, (1) detecton of the trajectory n admssblty, (2) the means of generatng alternatve mmedate ponts, and (3) the generated subgoal n forward order. he above algorthm can be modfed for use n the 3D path-plannng problem. 3D path plannng can be reduced to 2D path-plannng n a 2D path plane that passes ntal and fnal ponts. he 2D polygonal obstacle s vertces are obtaned from the ntersecton of the edges of obstacles and the 2D path plane. Fg. 10. Example of decson made by algorthm 2DAH. 5. CONCLUSIONS Computaton of the proposed mnmum drected dstance has been recast as the Mnkowsk sum of a two convex object problem. he proposed boundary-model approach can generate analytcal expressons for the Mnkowsk sum of two convex objects 2D; hence, MDLD can be expressed n analytcal form. For the 3D case, only a few edge-edge pars need be consdered durng computaton of the mnmum dstance between two convex polyhedra. he algorthm developed here can be used for collson detecton, computaton of the ds-
20 372 Fg. 11. A path planned by the 2DAH algorthm n 2D. tance between two polyhedra n three-dmensonal space, and robotcs path-plannng problems. he advantages of usng the Mnkowsk sum n computng dstance measures between two convex polyhedra are: (1) the measure of the separatng dstance or degree of penetraton can be consdered; (2) MDLE, MD and MDED functons can be computed n one routne; and (3) computaton complexty s lnear n 2D, and computaton s fast n 3D. REFERENCES 1. J.E. Bobrow, A drect mnmzaton approach for obtanng the dstance between convex polyhedra, Internatonal Journal of Robotcs Research, Vol. 8, No. 3, June 1989, pp C.E. Buckley, A foundaton for the flexble-trajectory approach to numerc path plannng, Internatonal Journal of Robotcs Research, Vol. 8, No. 3, June 1989, pp S.A. Cameron and R.K. Culley, Determnng the mnmum translatonal dstance between two convex polyhedra, IEEE Internatonal Conference on Robotcs and Automaton, 1986, pp F. Chn, and C.A. Wang, Optmal algorthms for the ntersecton and mnmum dstance problems between planar polygons, IEEE ransactons on Computers, Vol. C-32, 1983, pp E.G. Glbert, D.W. Johnson and S.S. Keerth, A fast procedure for computng the ds-
21 COMUING DISANCES BEWEEN CONVEX OLYHEDRA 373 tance between complex objects n three-dmensonal space, IEEE ransactons on Robotcs and Automaton, Vol. 4, No. 2, 1988, pp M.C. Ln and J. Canny, A fast algorthm for ncremental dstance calculaton, IEEE Internatonal Conference on Robotcs and Automaton, Aprl, 1991, pp Lozano-erez, Spatal plannng: a confguraton space approach, IEEE ransactons on Computers, Vol. 32, No. 2, 1983, pp V.J. Lumelsky, On fast computaton of dstance between lne segments, Informaton rocessng Letters, Vol. 21, August, 1985, pp W. Meyer, Dstances between boxes: applcatons to collson detecton and clppng, IEEE Internatonal Conference on Robotcs and Automaton, 1986, pp J.. Schwarz, Fndng the mnmum dstance between two convex polygons, Informaton rocessng Letters, Vol. 13, 1981, pp Chng-Long Shh (Iy )receved the B.S. and M.S. degrees from Natonal Chao ung Unversty, and the h. D. degree n electrcal engneerng from the Oho State Unversty n Currently, he s a professor n the Department of Electrcal Engnerng, Natonal awan Insttute of echnology. Hs research nterests nclude moble robot moton plannng and legged robot locomoton. Jane-Yu Lu (BØü)receved the B.S. degree n electronc engneerng from Chung-Yuan Chrstan Unversty n 1992 and the M.S. degree n electrcal engneerng from the Natonal awan Insttute of echnology n he subject of hs masters thess was mnmum drected dstance and path plannng n 2D space. He currently s employed by the Wnbond Electroncs Corp., workng on devce relabllty.
Parallelism for Nested Loops with Non-uniform and Flow Dependences
Parallelsm for Nested Loops wth Non-unform and Flow Dependences Sam-Jn Jeong Dept. of Informaton & Communcaton Engneerng, Cheonan Unversty, 5, Anseo-dong, Cheonan, Chungnam, 330-80, Korea. seong@cheonan.ac.kr
More informationRange images. Range image registration. Examples of sampling patterns. Range images and range surfaces
Range mages For many structured lght scanners, the range data forms a hghly regular pattern known as a range mage. he samplng pattern s determned by the specfc scanner. Range mage regstraton 1 Examples
More informationThe Greedy Method. Outline and Reading. Change Money Problem. Greedy Algorithms. Applications of the Greedy Strategy. The Greedy Method Technique
//00 :0 AM Outlne and Readng The Greedy Method The Greedy Method Technque (secton.) Fractonal Knapsack Problem (secton..) Task Schedulng (secton..) Mnmum Spannng Trees (secton.) Change Money Problem Greedy
More informationAn Optimal Algorithm for Prufer Codes *
J. Software Engneerng & Applcatons, 2009, 2: 111-115 do:10.4236/jsea.2009.22016 Publshed Onlne July 2009 (www.scrp.org/journal/jsea) An Optmal Algorthm for Prufer Codes * Xaodong Wang 1, 2, Le Wang 3,
More informationSubspace clustering. Clustering. Fundamental to all clustering techniques is the choice of distance measure between data points;
Subspace clusterng Clusterng Fundamental to all clusterng technques s the choce of dstance measure between data ponts; D q ( ) ( ) 2 x x = x x, j k = 1 k jk Squared Eucldean dstance Assumpton: All features
More information2x x l. Module 3: Element Properties Lecture 4: Lagrange and Serendipity Elements
Module 3: Element Propertes Lecture : Lagrange and Serendpty Elements 5 In last lecture note, the nterpolaton functons are derved on the bass of assumed polynomal from Pascal s trangle for the fled varable.
More informationProgramming in Fortran 90 : 2017/2018
Programmng n Fortran 90 : 2017/2018 Programmng n Fortran 90 : 2017/2018 Exercse 1 : Evaluaton of functon dependng on nput Wrte a program who evaluate the functon f (x,y) for any two user specfed values
More informationThe Shortest Path of Touring Lines given in the Plane
Send Orders for Reprnts to reprnts@benthamscence.ae 262 The Open Cybernetcs & Systemcs Journal, 2015, 9, 262-267 The Shortest Path of Tourng Lnes gven n the Plane Open Access Ljuan Wang 1,2, Dandan He
More informationSupport Vector Machines
/9/207 MIST.6060 Busness Intellgence and Data Mnng What are Support Vector Machnes? Support Vector Machnes Support Vector Machnes (SVMs) are supervsed learnng technques that analyze data and recognze patterns.
More informationHermite Splines in Lie Groups as Products of Geodesics
Hermte Splnes n Le Groups as Products of Geodescs Ethan Eade Updated May 28, 2017 1 Introducton 1.1 Goal Ths document defnes a curve n the Le group G parametrzed by tme and by structural parameters n the
More informationUNIT 2 : INEQUALITIES AND CONVEX SETS
UNT 2 : NEQUALTES AND CONVEX SETS ' Structure 2. ntroducton Objectves, nequaltes and ther Graphs Convex Sets and ther Geometry Noton of Convex Sets Extreme Ponts of Convex Set Hyper Planes and Half Spaces
More informationAn Application of the Dulmage-Mendelsohn Decomposition to Sparse Null Space Bases of Full Row Rank Matrices
Internatonal Mathematcal Forum, Vol 7, 2012, no 52, 2549-2554 An Applcaton of the Dulmage-Mendelsohn Decomposton to Sparse Null Space Bases of Full Row Rank Matrces Mostafa Khorramzadeh Department of Mathematcal
More informationR s s f. m y s. SPH3UW Unit 7.3 Spherical Concave Mirrors Page 1 of 12. Notes
SPH3UW Unt 7.3 Sphercal Concave Mrrors Page 1 of 1 Notes Physcs Tool box Concave Mrror If the reflectng surface takes place on the nner surface of the sphercal shape so that the centre of the mrror bulges
More informationAccounting for the Use of Different Length Scale Factors in x, y and z Directions
1 Accountng for the Use of Dfferent Length Scale Factors n x, y and z Drectons Taha Soch (taha.soch@kcl.ac.uk) Imagng Scences & Bomedcal Engneerng, Kng s College London, The Rayne Insttute, St Thomas Hosptal,
More information6.854 Advanced Algorithms Petar Maymounkov Problem Set 11 (November 23, 2005) With: Benjamin Rossman, Oren Weimann, and Pouya Kheradpour
6.854 Advanced Algorthms Petar Maymounkov Problem Set 11 (November 23, 2005) Wth: Benjamn Rossman, Oren Wemann, and Pouya Kheradpour Problem 1. We reduce vertex cover to MAX-SAT wth weghts, such that the
More informationLearning the Kernel Parameters in Kernel Minimum Distance Classifier
Learnng the Kernel Parameters n Kernel Mnmum Dstance Classfer Daoqang Zhang 1,, Songcan Chen and Zh-Hua Zhou 1* 1 Natonal Laboratory for Novel Software Technology Nanjng Unversty, Nanjng 193, Chna Department
More informationCompiler Design. Spring Register Allocation. Sample Exercises and Solutions. Prof. Pedro C. Diniz
Compler Desgn Sprng 2014 Regster Allocaton Sample Exercses and Solutons Prof. Pedro C. Dnz USC / Informaton Scences Insttute 4676 Admralty Way, Sute 1001 Marna del Rey, Calforna 90292 pedro@s.edu Regster
More informationAnalysis of Continuous Beams in General
Analyss of Contnuous Beams n General Contnuous beams consdered here are prsmatc, rgdly connected to each beam segment and supported at varous ponts along the beam. onts are selected at ponts of support,
More informationFast Computation of Shortest Path for Visiting Segments in the Plane
Send Orders for Reprnts to reprnts@benthamscence.ae 4 The Open Cybernetcs & Systemcs Journal, 04, 8, 4-9 Open Access Fast Computaton of Shortest Path for Vstng Segments n the Plane Ljuan Wang,, Bo Jang
More informationCMPS 10 Introduction to Computer Science Lecture Notes
CPS 0 Introducton to Computer Scence Lecture Notes Chapter : Algorthm Desgn How should we present algorthms? Natural languages lke Englsh, Spansh, or French whch are rch n nterpretaton and meanng are not
More informationLobachevsky State University of Nizhni Novgorod. Polyhedron. Quick Start Guide
Lobachevsky State Unversty of Nzhn Novgorod Polyhedron Quck Start Gude Nzhn Novgorod 2016 Contents Specfcaton of Polyhedron software... 3 Theoretcal background... 4 1. Interface of Polyhedron... 6 1.1.
More informationQuality Improvement Algorithm for Tetrahedral Mesh Based on Optimal Delaunay Triangulation
Intellgent Informaton Management, 013, 5, 191-195 Publshed Onlne November 013 (http://www.scrp.org/journal/m) http://dx.do.org/10.36/m.013.5601 Qualty Improvement Algorthm for Tetrahedral Mesh Based on
More informationRelated-Mode Attacks on CTR Encryption Mode
Internatonal Journal of Network Securty, Vol.4, No.3, PP.282 287, May 2007 282 Related-Mode Attacks on CTR Encrypton Mode Dayn Wang, Dongda Ln, and Wenlng Wu (Correspondng author: Dayn Wang) Key Laboratory
More informationA Binarization Algorithm specialized on Document Images and Photos
A Bnarzaton Algorthm specalzed on Document mages and Photos Ergna Kavalleratou Dept. of nformaton and Communcaton Systems Engneerng Unversty of the Aegean kavalleratou@aegean.gr Abstract n ths paper, a
More information3D vector computer graphics
3D vector computer graphcs Paolo Varagnolo: freelance engneer Padova Aprl 2016 Prvate Practce ----------------------------------- 1. Introducton Vector 3D model representaton n computer graphcs requres
More informationLECTURE : MANIFOLD LEARNING
LECTURE : MANIFOLD LEARNING Rta Osadchy Some sldes are due to L.Saul, V. C. Raykar, N. Verma Topcs PCA MDS IsoMap LLE EgenMaps Done! Dmensonalty Reducton Data representaton Inputs are real-valued vectors
More informationSum of Linear and Fractional Multiobjective Programming Problem under Fuzzy Rules Constraints
Australan Journal of Basc and Appled Scences, 2(4): 1204-1208, 2008 ISSN 1991-8178 Sum of Lnear and Fractonal Multobjectve Programmng Problem under Fuzzy Rules Constrants 1 2 Sanjay Jan and Kalash Lachhwan
More informationCluster Analysis of Electrical Behavior
Journal of Computer and Communcatons, 205, 3, 88-93 Publshed Onlne May 205 n ScRes. http://www.scrp.org/ournal/cc http://dx.do.org/0.4236/cc.205.350 Cluster Analyss of Electrcal Behavor Ln Lu Ln Lu, School
More informationOutline. Discriminative classifiers for image recognition. Where in the World? A nearest neighbor recognition example 4/14/2011. CS 376 Lecture 22 1
4/14/011 Outlne Dscrmnatve classfers for mage recognton Wednesday, Aprl 13 Krsten Grauman UT-Austn Last tme: wndow-based generc obect detecton basc ppelne face detecton wth boostng as case study Today:
More informationA SYSTOLIC APPROACH TO LOOP PARTITIONING AND MAPPING INTO FIXED SIZE DISTRIBUTED MEMORY ARCHITECTURES
A SYSOLIC APPROACH O LOOP PARIIONING AND MAPPING INO FIXED SIZE DISRIBUED MEMORY ARCHIECURES Ioanns Drosts, Nektaros Kozrs, George Papakonstantnou and Panayots sanakas Natonal echncal Unversty of Athens
More informationConstructing Minimum Connected Dominating Set: Algorithmic approach
Constructng Mnmum Connected Domnatng Set: Algorthmc approach G.N. Puroht and Usha Sharma Centre for Mathematcal Scences, Banasthal Unversty, Rajasthan 304022 usha.sharma94@yahoo.com Abstract: Connected
More informationA Fast Content-Based Multimedia Retrieval Technique Using Compressed Data
A Fast Content-Based Multmeda Retreval Technque Usng Compressed Data Borko Furht and Pornvt Saksobhavvat NSF Multmeda Laboratory Florda Atlantc Unversty, Boca Raton, Florda 3343 ABSTRACT In ths paper,
More informationDynamic wetting property investigation of AFM tips in micro/nanoscale
Dynamc wettng property nvestgaton of AFM tps n mcro/nanoscale The wettng propertes of AFM probe tps are of concern n AFM tp related force measurement, fabrcaton, and manpulaton technques, such as dp-pen
More informationModule 6: FEM for Plates and Shells Lecture 6: Finite Element Analysis of Shell
Module 6: FEM for Plates and Shells Lecture 6: Fnte Element Analyss of Shell 3 6.6. Introducton A shell s a curved surface, whch by vrtue of ther shape can wthstand both membrane and bendng forces. A shell
More informationLine Clipping by Convex and Nonconvex Polyhedra in E 3
Lne Clppng by Convex and Nonconvex Polyhedra n E 3 Václav Skala 1 Department of Informatcs and Computer Scence Unversty of West Bohema Unverztní 22, Box 314, 306 14 Plzeò Czech Republc e-mal: skala@kv.zcu.cz
More informationCourse Introduction. Algorithm 8/31/2017. COSC 320 Advanced Data Structures and Algorithms. COSC 320 Advanced Data Structures and Algorithms
Course Introducton Course Topcs Exams, abs, Proects A quc loo at a few algorthms 1 Advanced Data Structures and Algorthms Descrpton: We are gong to dscuss algorthm complexty analyss, algorthm desgn technques
More informationAccessibility Analysis for the Automatic Contact and Non-contact Inspection on Coordinate Measuring Machines
Proceedngs of the World Congress on Engneerng 008 Vol I Accessblty Analyss for the Automatc Contact and Non-contact Inspecton on Coordnate Measurng Machnes B. J. Álvarez, P. Fernández, J. C. Rco and G.
More informationHigh-Boost Mesh Filtering for 3-D Shape Enhancement
Hgh-Boost Mesh Flterng for 3-D Shape Enhancement Hrokazu Yagou Λ Alexander Belyaev y Damng We z Λ y z ; ; Shape Modelng Laboratory, Unversty of Azu, Azu-Wakamatsu 965-8580 Japan y Computer Graphcs Group,
More informationNUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS
ARPN Journal of Engneerng and Appled Scences 006-017 Asan Research Publshng Network (ARPN). All rghts reserved. NUMERICAL SOLVING OPTIMAL CONTROL PROBLEMS BY THE METHOD OF VARIATIONS Igor Grgoryev, Svetlana
More informationTN348: Openlab Module - Colocalization
TN348: Openlab Module - Colocalzaton Topc The Colocalzaton module provdes the faclty to vsualze and quantfy colocalzaton between pars of mages. The Colocalzaton wndow contans a prevew of the two mages
More informationSequential search. Building Java Programs Chapter 13. Sequential search. Sequential search
Sequental search Buldng Java Programs Chapter 13 Searchng and Sortng sequental search: Locates a target value n an array/lst by examnng each element from start to fnsh. How many elements wll t need to
More informationMULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION
MULTISPECTRAL IMAGES CLASSIFICATION BASED ON KLT AND ATR AUTOMATIC TARGET RECOGNITION Paulo Quntlano 1 & Antono Santa-Rosa 1 Federal Polce Department, Brasla, Brazl. E-mals: quntlano.pqs@dpf.gov.br and
More informationMathematics 256 a course in differential equations for engineering students
Mathematcs 56 a course n dfferental equatons for engneerng students Chapter 5. More effcent methods of numercal soluton Euler s method s qute neffcent. Because the error s essentally proportonal to the
More informationContent Based Image Retrieval Using 2-D Discrete Wavelet with Texture Feature with Different Classifiers
IOSR Journal of Electroncs and Communcaton Engneerng (IOSR-JECE) e-issn: 78-834,p- ISSN: 78-8735.Volume 9, Issue, Ver. IV (Mar - Apr. 04), PP 0-07 Content Based Image Retreval Usng -D Dscrete Wavelet wth
More informationActive Contours/Snakes
Actve Contours/Snakes Erkut Erdem Acknowledgement: The sldes are adapted from the sldes prepared by K. Grauman of Unversty of Texas at Austn Fttng: Edges vs. boundares Edges useful sgnal to ndcate occludng
More informationBrave New World Pseudocode Reference
Brave New World Pseudocode Reference Pseudocode s a way to descrbe how to accomplsh tasks usng basc steps lke those a computer mght perform. In ths week s lab, you'll see how a form of pseudocode can be
More informationThe Codesign Challenge
ECE 4530 Codesgn Challenge Fall 2007 Hardware/Software Codesgn The Codesgn Challenge Objectves In the codesgn challenge, your task s to accelerate a gven software reference mplementaton as fast as possble.
More informationPositive Semi-definite Programming Localization in Wireless Sensor Networks
Postve Sem-defnte Programmng Localzaton n Wreless Sensor etworks Shengdong Xe 1,, Jn Wang, Aqun Hu 1, Yunl Gu, Jang Xu, 1 School of Informaton Scence and Engneerng, Southeast Unversty, 10096, anjng Computer
More informationChapter 6 Programmng the fnte element method Inow turn to the man subject of ths book: The mplementaton of the fnte element algorthm n computer programs. In order to make my dscusson as straghtforward
More informationMachine Learning: Algorithms and Applications
14/05/1 Machne Learnng: Algorthms and Applcatons Florano Zn Free Unversty of Bozen-Bolzano Faculty of Computer Scence Academc Year 011-01 Lecture 10: 14 May 01 Unsupervsed Learnng cont Sldes courtesy of
More informationAnnouncements. Supervised Learning
Announcements See Chapter 5 of Duda, Hart, and Stork. Tutoral by Burge lnked to on web page. Supervsed Learnng Classfcaton wth labeled eamples. Images vectors n hgh-d space. Supervsed Learnng Labeled eamples
More information12/2/2009. Announcements. Parametric / Non-parametric. Case-Based Reasoning. Nearest-Neighbor on Images. Nearest-Neighbor Classification
Introducton to Artfcal Intellgence V22.0472-001 Fall 2009 Lecture 24: Nearest-Neghbors & Support Vector Machnes Rob Fergus Dept of Computer Scence, Courant Insttute, NYU Sldes from Danel Yeung, John DeNero
More informationA Computational Feasibility Study of Failure-Tolerant Path Planning
A Computatonal Feasblty Study of Falure-Tolerant Path Plannng Rodrgo S. Jamsola, Jr., Anthony A. Macejewsk, & Rodney G. Roberts Colorado State Unversty, Ft. Collns, Colorado 80523-1373, USA Florda A&M-Florda
More informationCollision Detection. Overview. Efficient Collision Detection. Collision Detection with Rays: Example. C = nm + (n choose 2)
Overvew Collson detecton wth Rays Collson detecton usng BSP trees Herarchcal Collson Detecton OBB tree, k-dop tree algorthms Multple object CD system Collson Detecton Fundamental to graphcs, VR applcatons
More informationPolyhedral Compilation Foundations
Polyhedral Complaton Foundatons Lous-Noël Pouchet pouchet@cse.oho-state.edu Dept. of Computer Scence and Engneerng, the Oho State Unversty Feb 8, 200 888., Class # Introducton: Polyhedral Complaton Foundatons
More informationSkew Angle Estimation and Correction of Hand Written, Textual and Large areas of Non-Textual Document Images: A Novel Approach
Angle Estmaton and Correcton of Hand Wrtten, Textual and Large areas of Non-Textual Document Images: A Novel Approach D.R.Ramesh Babu Pyush M Kumat Mahesh D Dhannawat PES Insttute of Technology Research
More informationMachine Learning. Support Vector Machines. (contains material adapted from talks by Constantin F. Aliferis & Ioannis Tsamardinos, and Martin Law)
Machne Learnng Support Vector Machnes (contans materal adapted from talks by Constantn F. Alfers & Ioanns Tsamardnos, and Martn Law) Bryan Pardo, Machne Learnng: EECS 349 Fall 2014 Support Vector Machnes
More informationFor instance, ; the five basic number-sets are increasingly more n A B & B A A = B (1)
Secton 1.2 Subsets and the Boolean operatons on sets If every element of the set A s an element of the set B, we say that A s a subset of B, or that A s contaned n B, or that B contans A, and we wrte A
More informationKiran Joy, International Journal of Advanced Engineering Technology E-ISSN
Kran oy, nternatonal ournal of Advanced Engneerng Technology E-SS 0976-3945 nt Adv Engg Tech/Vol. V/ssue /Aprl-une,04/9-95 Research Paper DETERMATO O RADATVE VEW ACTOR WTOUT COSDERG TE SADOWG EECT Kran
More informationReading. 14. Subdivision curves. Recommended:
eadng ecommended: Stollntz, Deose, and Salesn. Wavelets for Computer Graphcs: heory and Applcatons, 996, secton 6.-6., A.5. 4. Subdvson curves Note: there s an error n Stollntz, et al., secton A.5. Equaton
More informationSteps for Computing the Dissimilarity, Entropy, Herfindahl-Hirschman and. Accessibility (Gravity with Competition) Indices
Steps for Computng the Dssmlarty, Entropy, Herfndahl-Hrschman and Accessblty (Gravty wth Competton) Indces I. Dssmlarty Index Measurement: The followng formula can be used to measure the evenness between
More informationFEATURE EXTRACTION. Dr. K.Vijayarekha. Associate Dean School of Electrical and Electronics Engineering SASTRA University, Thanjavur
FEATURE EXTRACTION Dr. K.Vjayarekha Assocate Dean School of Electrcal and Electroncs Engneerng SASTRA Unversty, Thanjavur613 41 Jont Intatve of IITs and IISc Funded by MHRD Page 1 of 8 Table of Contents
More informationClassifier Selection Based on Data Complexity Measures *
Classfer Selecton Based on Data Complexty Measures * Edth Hernández-Reyes, J.A. Carrasco-Ochoa, and J.Fco. Martínez-Trndad Natonal Insttute for Astrophyscs, Optcs and Electroncs, Lus Enrque Erro No.1 Sta.
More informationROBOT KINEMATICS. ME Robotics ME Robotics
ROBOT KINEMATICS Purpose: The purpose of ths chapter s to ntroduce you to robot knematcs, and the concepts related to both open and closed knematcs chans. Forward knematcs s dstngushed from nverse knematcs.
More informationData Representation in Digital Design, a Single Conversion Equation and a Formal Languages Approach
Data Representaton n Dgtal Desgn, a Sngle Converson Equaton and a Formal Languages Approach Hassan Farhat Unversty of Nebraska at Omaha Abstract- In the study of data representaton n dgtal desgn and computer
More informationA Robust Method for Estimating the Fundamental Matrix
Proc. VIIth Dgtal Image Computng: Technques and Applcatons, Sun C., Talbot H., Ourseln S. and Adraansen T. (Eds.), 0- Dec. 003, Sydney A Robust Method for Estmatng the Fundamental Matrx C.L. Feng and Y.S.
More informationREFRACTION. a. To study the refraction of light from plane surfaces. b. To determine the index of refraction for Acrylic and Water.
Purpose Theory REFRACTION a. To study the refracton of lght from plane surfaces. b. To determne the ndex of refracton for Acrylc and Water. When a ray of lght passes from one medum nto another one of dfferent
More informationImprovement of Spatial Resolution Using BlockMatching Based Motion Estimation and Frame. Integration
Improvement of Spatal Resoluton Usng BlockMatchng Based Moton Estmaton and Frame Integraton Danya Suga and Takayuk Hamamoto Graduate School of Engneerng, Tokyo Unversty of Scence, 6-3-1, Nuku, Katsuska-ku,
More informationF Geometric Mean Graphs
Avalable at http://pvamu.edu/aam Appl. Appl. Math. ISSN: 1932-9466 Vol. 10, Issue 2 (December 2015), pp. 937-952 Applcatons and Appled Mathematcs: An Internatonal Journal (AAM) F Geometrc Mean Graphs A.
More informationArray transposition in CUDA shared memory
Array transposton n CUDA shared memory Mke Gles February 19, 2014 Abstract Ths short note s nspred by some code wrtten by Jeremy Appleyard for the transposton of data through shared memory. I had some
More informationImage Alignment CSC 767
Image Algnment CSC 767 Image algnment Image from http://graphcs.cs.cmu.edu/courses/15-463/2010_fall/ Image algnment: Applcatons Panorama sttchng Image algnment: Applcatons Recognton of object nstances
More informationNon-Split Restrained Dominating Set of an Interval Graph Using an Algorithm
Internatonal Journal of Advancements n Research & Technology, Volume, Issue, July- ISS - on-splt Restraned Domnatng Set of an Interval Graph Usng an Algorthm ABSTRACT Dr.A.Sudhakaraah *, E. Gnana Deepka,
More informationProblem Definitions and Evaluation Criteria for Computational Expensive Optimization
Problem efntons and Evaluaton Crtera for Computatonal Expensve Optmzaton B. Lu 1, Q. Chen and Q. Zhang 3, J. J. Lang 4, P. N. Suganthan, B. Y. Qu 6 1 epartment of Computng, Glyndwr Unversty, UK Faclty
More informationPrivate Information Retrieval (PIR)
2 Levente Buttyán Problem formulaton Alce wants to obtan nformaton from a database, but she does not want the database to learn whch nformaton she wanted e.g., Alce s an nvestor queryng a stock-market
More informationProblem Set 3 Solutions
Introducton to Algorthms October 4, 2002 Massachusetts Insttute of Technology 6046J/18410J Professors Erk Demane and Shaf Goldwasser Handout 14 Problem Set 3 Solutons (Exercses were not to be turned n,
More informationAn Image Fusion Approach Based on Segmentation Region
Rong Wang, L-Qun Gao, Shu Yang, Yu-Hua Cha, and Yan-Chun Lu An Image Fuson Approach Based On Segmentaton Regon An Image Fuson Approach Based on Segmentaton Regon Rong Wang, L-Qun Gao, Shu Yang 3, Yu-Hua
More informationLoad Balancing for Hex-Cell Interconnection Network
Int. J. Communcatons, Network and System Scences,,, - Publshed Onlne Aprl n ScRes. http://www.scrp.org/journal/jcns http://dx.do.org/./jcns.. Load Balancng for Hex-Cell Interconnecton Network Saher Manaseer,
More informationConcurrent Apriori Data Mining Algorithms
Concurrent Apror Data Mnng Algorthms Vassl Halatchev Department of Electrcal Engneerng and Computer Scence York Unversty, Toronto October 8, 2015 Outlne Why t s mportant Introducton to Assocaton Rule Mnng
More informationAn Improved Image Segmentation Algorithm Based on the Otsu Method
3th ACIS Internatonal Conference on Software Engneerng, Artfcal Intellgence, Networkng arallel/dstrbuted Computng An Improved Image Segmentaton Algorthm Based on the Otsu Method Mengxng Huang, enjao Yu,
More informationEcient Computation of the Most Probable Motion from Fuzzy. Moshe Ben-Ezra Shmuel Peleg Michael Werman. The Hebrew University of Jerusalem
Ecent Computaton of the Most Probable Moton from Fuzzy Correspondences Moshe Ben-Ezra Shmuel Peleg Mchael Werman Insttute of Computer Scence The Hebrew Unversty of Jerusalem 91904 Jerusalem, Israel Emal:
More informationCE 221 Data Structures and Algorithms
CE 1 ata Structures and Algorthms Chapter 4: Trees BST Text: Read Wess, 4.3 Izmr Unversty of Economcs 1 The Search Tree AT Bnary Search Trees An mportant applcaton of bnary trees s n searchng. Let us assume
More informationLecture #15 Lecture Notes
Lecture #15 Lecture Notes The ocean water column s very much a 3-D spatal entt and we need to represent that structure n an economcal way to deal wth t n calculatons. We wll dscuss one way to do so, emprcal
More informationComputer Animation and Visualisation. Lecture 4. Rigging / Skinning
Computer Anmaton and Vsualsaton Lecture 4. Rggng / Sknnng Taku Komura Overvew Sknnng / Rggng Background knowledge Lnear Blendng How to decde weghts? Example-based Method Anatomcal models Sknnng Assume
More informationSolving Route Planning Using Euler Path Transform
Solvng Route Plannng Usng Euler Path ransform Y-Chong Zeng Insttute of Informaton Scence Academa Snca awan ychongzeng@s.snca.edu.tw Abstract hs paper presents a method to solve route plannng problem n
More informationKent State University CS 4/ Design and Analysis of Algorithms. Dept. of Math & Computer Science LECT-16. Dynamic Programming
CS 4/560 Desgn and Analyss of Algorthms Kent State Unversty Dept. of Math & Computer Scence LECT-6 Dynamc Programmng 2 Dynamc Programmng Dynamc Programmng, lke the dvde-and-conquer method, solves problems
More informationAn Iterative Solution Approach to Process Plant Layout using Mixed Integer Optimisation
17 th European Symposum on Computer Aded Process Engneerng ESCAPE17 V. Plesu and P.S. Agach (Edtors) 2007 Elsever B.V. All rghts reserved. 1 An Iteratve Soluton Approach to Process Plant Layout usng Mxed
More informationParallel matrix-vector multiplication
Appendx A Parallel matrx-vector multplcaton The reduced transton matrx of the three-dmensonal cage model for gel electrophoress, descrbed n secton 3.2, becomes excessvely large for polymer lengths more
More informationEdge Detection in Noisy Images Using the Support Vector Machines
Edge Detecton n Nosy Images Usng the Support Vector Machnes Hlaro Gómez-Moreno, Saturnno Maldonado-Bascón, Francsco López-Ferreras Sgnal Theory and Communcatons Department. Unversty of Alcalá Crta. Madrd-Barcelona
More informationCS1100 Introduction to Programming
Factoral (n) Recursve Program fact(n) = n*fact(n-) CS00 Introducton to Programmng Recurson and Sortng Madhu Mutyam Department of Computer Scence and Engneerng Indan Insttute of Technology Madras nt fact
More informationAssignment # 2. Farrukh Jabeen Algorithms 510 Assignment #2 Due Date: June 15, 2009.
Farrukh Jabeen Algorthms 51 Assgnment #2 Due Date: June 15, 29. Assgnment # 2 Chapter 3 Dscrete Fourer Transforms Implement the FFT for the DFT. Descrbed n sectons 3.1 and 3.2. Delverables: 1. Concse descrpton
More informationSketching of Mirror-symmetric Shapes
IEEE TRANSACTIONS ON TRANSACTIONS ON VISUALIZATION AND COMPUTER GRAPHICS Sketchng of Mrror-symmetrc Shapes Frederc Corder Hyewon Seo Jnho Park and Junyong Noh Fg.. Sketchng of a symmetrc shape. Abstract
More informationAn Approach in Coloring Semi-Regular Tilings on the Hyperbolic Plane
An Approach n Colorng Sem-Regular Tlngs on the Hyperbolc Plane Ma Louse Antonette N De Las Peñas, mlp@mathscmathadmueduph Glenn R Lago, glago@yahoocom Math Department, Ateneo de Manla Unversty, Loyola
More informationOpen Access A New Algorithm for the Shortest Path of Touring Disjoint Convex Polygons
Send Orders for Reprnts to reprnts@benthamscence.ae 1364 The Open Automaton and Control Systems Journal, 2015, 7, 1364-1368 Open Access A New Algorthm for the Shortest Path of Tourng Dsjont Convex Polygons
More informationSolving two-person zero-sum game by Matlab
Appled Mechancs and Materals Onlne: 2011-02-02 ISSN: 1662-7482, Vols. 50-51, pp 262-265 do:10.4028/www.scentfc.net/amm.50-51.262 2011 Trans Tech Publcatons, Swtzerland Solvng two-person zero-sum game by
More informationA Clustering Algorithm for Chinese Adjectives and Nouns 1
Clusterng lgorthm for Chnese dectves and ouns Yang Wen, Chunfa Yuan, Changnng Huang 2 State Key aboratory of Intellgent Technology and System Deptartment of Computer Scence & Technology, Tsnghua Unversty,
More informationModule Management Tool in Software Development Organizations
Journal of Computer Scence (5): 8-, 7 ISSN 59-66 7 Scence Publcatons Management Tool n Software Development Organzatons Ahmad A. Al-Rababah and Mohammad A. Al-Rababah Faculty of IT, Al-Ahlyyah Amman Unversty,
More informationSLAM Summer School 2006 Practical 2: SLAM using Monocular Vision
SLAM Summer School 2006 Practcal 2: SLAM usng Monocular Vson Javer Cvera, Unversty of Zaragoza Andrew J. Davson, Imperal College London J.M.M Montel, Unversty of Zaragoza. josemar@unzar.es, jcvera@unzar.es,
More informationCollaboratively Regularized Nearest Points for Set Based Recognition
Academc Center for Computng and Meda Studes, Kyoto Unversty Collaboratvely Regularzed Nearest Ponts for Set Based Recognton Yang Wu, Mchhko Mnoh, Masayuk Mukunok Kyoto Unversty 9/1/013 BMVC 013 @ Brstol,
More informationRamsey numbers of cubes versus cliques
Ramsey numbers of cubes versus clques Davd Conlon Jacob Fox Choongbum Lee Benny Sudakov Abstract The cube graph Q n s the skeleton of the n-dmensonal cube. It s an n-regular graph on 2 n vertces. The Ramsey
More informationGSLM Operations Research II Fall 13/14
GSLM 58 Operatons Research II Fall /4 6. Separable Programmng Consder a general NLP mn f(x) s.t. g j (x) b j j =. m. Defnton 6.. The NLP s a separable program f ts objectve functon and all constrants are
More information