Gen. Math. Notes, Vol. 24, No. 1, September 2014, pp. 98-108 ISSN 2219-7184; Copyright © ICSRS Publication, 2014 www.i-csrs.org Available free online at http://www.geman.in On Characterization of Inextensible Flows of Dual Curves According to Dual Darboux Frame Zehra Ekinci Celal Bayar University, Faculty of Arts and Sciences Department of Mathematics, Muradiye Campus Muradiye, Manisa, Turkey E-mail: zehra.ari@cbu.edu.tr (Received: 30-5-14 / Accepted: 12-7-14) Abstract In this paper, we study inextensible flows dual curves according to dual Darboux frame. Necessary and sufficient conditions for an inelastic dual curve flow are expressed as a partial differential equation involving the dual geodesic curvature. Keywords: Dual Darboux Frame, Inextensible flows. 1 Introduction Recently, the study of the motion of inelastic curves has an important role. The time evolution of a curve represented by its corresponding flow. The flow of a curve is said to be inextensible if, firstly its arc length is preserved and secondly its intrinsic curvature is preserved. Physically, inextensible curve flows give rise to motions in which no strain energy is induced. The swinging motion of cord of fixed length, for example, or of a piece of paper carried by the wind, can be On Characterization of Inextensible Flows of... 99 described by inextensible curve. Some movement in nature is inspired to examine flow of curves as snake and elephant's trunk movement. For example, both Chirikjian and Burdick [5] and Mochiyama et al. [9] study the shape control hyper-redundant, or snake-like robots. Inextensible curve and surface flows emerge many problems in computer vision [14] and computer animation [10]. Particularly, inextensible time evolution of curves and surfaces is examined mathematically. Significant methods of this article developed by Gage and Hamilton [11], and Grayson [13] for studying the shrinking of closed plane curves to a circle via the heat equation. In [12] Gage also studies area preserving evolution of inelastic plane curves. Another related study is that [6], which considers less restrictive mappings that locally preserve volume only. In [2, 3] Know et al. study evolution of inelastic plane curves, and inextensible flows of curves and developable surfaces. In this paper, we study inextensible flows dual curves according to dual Darboux frame. Necessary and sufficient conditions for an inelastic dual curve flow are expressed as a partial differential equation involving the dual geodesic curvature. 2 Preliminaries Let D = IR × IR = {a = ( a, a ∗ ) : a, a ∗ ∈ IR} be the set of the pairs (a, a ∗ ) . For a = (a, a ∗ ) , b = (b, b∗ ) ∈ D the following operations are defined on D: Equality: Addition: Multiplication: a = b ⇔ a = b, a ∗ = b ∗ a + b ⇔ ( a + b, a ∗ + b ∗ ) ab ⇔ (ab, ab∗ + a ∗b) The element ε = (0,1) ∈ D satisfies the relationships ε ≠ 0 , ε 2 = 0 , ε 1 = 1ε = ε (2.1) Let consider the element a ∈ D of the form a = (a, 0) . Then the mapping f : D → IR, f (a, 0) = a is a isomorphism. So, we can write a = (a, 0) . By the multiplication rule we have that a = (a, a ∗ ) = a + ε a∗ Then a = a + ε a∗ is called dual number and ε is called dual unit. Thus the set of dual numbers is given by (2.2) D = {a = a + ε a ∗ : a, a ∗ ∈ IR, ε 2 = 0} 100 Zehra Ekinci The set D forms a commutative group under addition. The associative laws hold for multiplication. Dual numbers are distributive and form a ring over the real number field [4]. Dual function of dual number presents a mapping of a dual numbers space on itself. Properties of dual functions were thoroughly investigated by Dimentberg [4]. He derived the general expression for dual analytic (differentiable) function as follows f ( x ) = f ( x + ε x∗ ) = f ( x) + ε x∗ f ′( x) , (2.3) where f ′( x ) is derivative of f ( x ) and x, x∗ ∈ IR . This definition allows us to write the dual forms of some well-known functions as follows cos( x ) = cos( x + ε x∗ ) = cos( x) − ε x∗ sin( x), ∗ ∗ sin( x ) = sin( x + ε x ) = sin( x) + ε x cos( x), ∗ x = x + ε x∗ = x + ε x , ( x > 0). 2 x (2.4) Let D 3 = D × D × D be the set of all triples of dual numbers, i.e., D3 = {aɶ = (a1 , a2 , a3 ) : ai ∈ D, i = 1, 2,3} (2.5) Then the set D 3 is called dual space. The elements of D 3 are called dual vectors. Similar to the dual numbers, a dual vector aɶ may be expressed in the form aɶ = a + ε a ∗ = (a , a ∗ ) , where a and a ∗ are the vectors of IR 3 . Then for any vectors aɶ = a + ε a ∗ and bɶ = b + ε b ∗ of D 3 , the scalar product and the vector product are defined by (2.6) aɶ , bɶ = a , b + ε a , b ∗ + a ∗ , b , ( and ) ( ) aɶ × bɶ = a × b + ε a × b * + a ∗ × b , (2.7) Respectively, where a , b and a × b are the inner product and the vector product of the vectors a and a ∗ in IR 3 , respectively. The norm of a dual vector aɶ is given by aɶ = a + ε a, a∗ a , (a ≠ 0) . (2.8) On Characterization of Inextensible Flows of... 101 A dual vector aɶ with norm 1 + ε 0 is called dual unit vector. The set of dual unit vectors is given by Sɶ 2 = {aɶ = ( a1 , a2 , a3 ) ∈ D 3 : aɶ , aɶ = 1 + ε 0} , (2.9) and called dual unit sphere[8,15]. In the Euclidean 3-space IR3 , an oriented line L is determined by a point p ∈ L and a unit vector a . Then, one can define a ∗ = p × a which is called moment vector. The value of a ∗ does not depend on the point p , because any other point q in L can be given by q = p + λ a and then a ∗ = p × a = q × a . Reciprocally, when such a pair (a , a ∗ ) is given, one recovers the line L as L = {( a × a ∗ ) + λ a : a , a ∗ ∈ E 3 , λ ∈ IR } , written in parametric equations. The vectors a and a ∗ are not independent of one another and they satisfy the following relationships a , a = 1, a , a ∗ = 0 (2.10) The components ai , ai∗ (1 ≤ i ≤ 3) of the vectors a and a ∗ are called the normalized Plucker coordinates of the line L . We see that the dual unit vector aɶ = a + ε a ∗ corresponds to the line L . This correspondence is known as E. Study Mapping: There exists a one-to-one correspondence between the vectors of dual unit sphere Sɶ 2 and the directed lines of the space IR 3 . By the aid of this correspondence, the properties of the spatial motion of a line can be derived. Hence, the geometry of ruled surface is represented by the geometry of dual curves lying on the dual unit sphere Sɶ 2 . The angle θ = θ + εθ * between two dual unit vectors aɶ , bɶ is called dual angle and defined by (2.11) aɶ , bɶ = cos θ = cos θ − εθ * sin θ . By considering the E. Study Mapping, the geometric interpretation of dual angle is that, θ is the real angle between the lines L1 , L2 corresponding to the dual unit vectors aɶ , bɶ , respectively, and θ * is the shortest distance between those lines [9]. Let now ( xɶ ) be a dual curve represented by the dual vector eɶ (u ) = e (u ) + ε e ∗ (u ) . The unit vector e draws a curve on the real unit sphere S 2 and is called the (real) indicatrix of ( xɶ ) . We suppose throughout that it is not a single point. We take the parameter u as the arc-length parameter s of the real indicatrix and denote the differentiation with respect to s by primes. Then we have e′, e′ = 1 . The vector 102 Zehra Ekinci e′ = t is the unit vector parallel to the tangent of the indicatrix. The equation e ∗ ( s ) = p ( s ) × e ( s ) has infinity of solutions for the function p ( s ) . If we take po ( s ) as a solution, the set of all solutions is given by p ( s ) = po ( s ) + λ ( s ) e ( s ) , where λ is a real scalar function of s . Therefore we have p′, e′ = po′ , e′ + λ . By taking λ = λo = − po′ , e′ we see that po ( s ) + λo ( s ) e ( s ) = c ( s ) is the unique solution for p ( s ) with c′, e′ = 0 . Then, the given dual curve ( xɶ ) corresponding to the ruled surface ϕe = c ( s ) + ve ( s ) (2.12) may be represented by eɶ ( s ) = e + ε c × e (2.13) where e , e = 1 , e′, e′ = 1 , c′, e′ = 0 . (2.14) Then we have eɶ′ = t + ε det(c′, e , t ) = 1 + ε∆ (2.15) where ∆ = det(c′, e , t ) . The dual arc-length s of the dual curve ( xɶ ) is given by s s s 0 0 0 s = ∫ eɶ′(u ) du = ∫ (1 + ε∆)du = s + ε ∫ ∆du (2.16) From (16) we have s′ = 1 + ε∆ . Therefore, the dual unit tangent to the curve eɶ ( s ) is given by deɶ eɶ′ eɶ′ = = = tɶ = t + ε (c × t ) (2.17) ds s′ 1 + ε∆ Introducing the dual unit vector gɶ = eɶ × tɶ = g + ε c × g we have the dual frame {eɶ, tɶ, gɶ} which is known as dual geodesic trihedron or dual Darboux frame of ϕe (or (eɶ ) ). Also, it is well known that the real orthonormal frame {e , t , g} is called the geodesic trihedron of the indicatrix e ( s ) with the derivations e′ = t , t ′ = γ g − e , g ′ = −γ t (2.18) where γ is called the conical curvature [1]. Similar to (2.18), the derivatives of the vectors of the dual frame {eɶ, tɶ, gɶ } are given by deɶ ɶ dtɶ dgɶ =t, = γ gɶ − eɶ, = −γ tɶ ds ds ds (2.19) γ = γ + ε (δ − γ∆) , δ = c′, e (2.20) where On Characterization of Inextensible Flows of... 103 and the dual darboux vector of the frame is dɶ = γ eɶ + gɶ . From the definition of ∆ and (2.20) we also have c′ = δ e + ∆g (2.21) The dual curvature of dual curve (ruled surface) eɶ ( s ) is R= 1 (2.22) (1 + γ 2 ) The unit vector dɶo with the same sense as the Darboux vector dɶ = γ eɶ + gɶ is given by γ 1 dɶo = eɶ + gɶ (2.23) (1 + γ 2 ) (1 + γ 2 ) Then, the dual angle between dɶo and eɶ satisfies the followings cos ρ = γ (1 + γ 2 ) , sin ρ = 1 (1 + γ 2 ) (2.24) where ρ is the dual spherical radius of curvature. Hence R = sin ρ , γ = cot ρ [7]. 3 Inextensible Flows of Dual Curves according to Dual Darboux Frame Throughout this study, we assume that xɶ :[0,1] × [0, ω ] → D3 is a one parameter family of smooth dual curves in dual space D3 . Let u parametrization variable, 0 ≤ u ≤ 1 . be the curve The arclenght of xɶ is given by u s =∫ 0 ∂xɶ du = ∫ v du ∂u 0 u (3.1) where ∂xɶ ∂xɶ ∂xɶ = , ∂u ∂u ∂u 12 . and we assume that s ∗ is a parameter in terms of s . The operator terms of u by ∂ 1 ∂ = ∂s v ∂u (3.2) ∂ is given in ∂s 104 Zehra Ekinci where v= ∂xɶ . ∂u The arclenght parameter is ds = v du . Any flow of xɶ can be represented as ∂xɶ = f1eɶ + f 2tɶ + f 3 gɶ . ∂t (3.3) The arclength is given up to a constant by u s (u , t ) = ∫ v du . 0 In the dual space the requirement that the curve not be exposed to any elongation or compression can be expressed by the condition ∂ ∂v s (u , t ) = ∫ du = 0 ∂t ∂t 0 u (3.4) for u ∈ [0,1] . Definition 3.1: A dual curve evolution xɶ (u , t ) and ∂xɶ flow of this curve in D3 ∂t are said to be inextensible if ∂ ∂xɶ =0. ∂ t ∂u ∂xɶ be a smooth flow of the dual curve xɶ . The flow is ∂u inextensible if and only if Lemma 3.2: Let ∂v ∂f 2 = + f1v − γ f3v ∂t ∂u ∂f 2∗ = γ f 3∗v + γ ∗ f3v + γ f3v∗ − f1∗v − f1 v∗ . ∂u Proof: Suppose that (3.5) ∂xɶ be a smooth flow of the curve xɶ . Considering definition ∂u of xɶ , we have v2 = ∂xɶ ∂xɶ , ∂u ∂u (3.6) On Characterization of Inextensible Flows of... 105 ∂ ∂ and commute since and are independent coordinates. Then, by ∂u ∂t differentiating of formula (3.6) we get 2v On the other hand, changing ∂v ∂ = ∂t ∂t ∂xɶ ∂xɶ . , ∂u ∂u ∂ ∂ and we ∂u ∂t v ∂v ∂xɶ ∂ ∂xɶ = , ∂t ∂u ∂u ∂ t From equation (3.3), we obtain v ∂v ∂xɶ ∂ = , ( f1eɶ + f 2tɶ + f3 gɶ ) . ∂t ∂u ∂u By the formula of dual darboux, we have ∂f ∂f ∂f ∂v = tɶ, 1 − f 2 v eɶ + f1v + 2 − γ f 3v tɶ + 3 + f 2γ v gɶ . ∂t ∂u ∂u ∂u If above equation is separated real and dual part then we have ∂v ∂f = f1v + 2 − γ f 3v ∂t ∂u and ∂v∗ ∂f 2∗ = − γ f 3∗v − γ ∗ f 3v − γ f 3v∗ + f1∗v + f1 v∗ . ∂t ∂u respectively. Desired expression is obtained with definition 3.1. ∂xɶ = f1eɶ + f 2tɶ + f 3 gɶ be a smooth flow of the curve xɶ . The flow ∂u is inextensible if and only if ∂f 2 = γ f 3 + f1 ∂s (3.7) ∂f 2∗ ∗ ∗ ∗ = γ f 3 + γ f 3 + f1 ∂s ∂xɶ Proof: Let be extensible. From Eq. (3.4), we have ∂u Theorem 3.3: Let 106 Zehra Ekinci ∂ ∂v ∂f s (u , t ) = ∫ du = ∫ ( 2 + f1v − γ f 3v )du = 0 ∂t ∂t ∂u 0 0 u u (3.8) ∀u ∈ [0,1] . Substituting (3.5) in (3.8) and this expression separating dual and real part complete the proof of the theorem. We suppose that v = 1 and the local coordinate u corresponds to the curve arc length s . Now we give following lemma that necessary. Lemma 3.4: ∂f ∂eɶ ∂f1 = − f 2 eɶ + f 2γ + 3 gɶ , ∂ t ∂s ∂s where ψ = (3.9) ∂tɶ = −eɶ + ψ gɶ , ∂t (3.10) ∂f ∂gɶ = − f 2γ + 3 eɶ −ψ tɶ , ∂t ∂s (3.11) ∂tɶ , gɶ . ∂t Proof: Considering definition of xɶ , we have ∂eɶ ∂ ∂xɶ ∂ = = ( f1eɶ + f 2tɶ + f 3 gɶ ). ∂ t ∂ t ∂s ∂s Using the Darboux equations, we get ∂f ∂eɶ ∂f1 ∂f = − f 2 eɶ + f1 + 2 − γ f 3 tɶ + f 2γ + 3 gɶ . ∂ t ∂s ∂s ∂s Substituting (3.7) in (3.12), we have ∂f ∂eɶ ∂f1 = − f 2 eɶ + f 2γ + 3 gɶ . ∂ t ∂s ∂s Now differentiate the dual Darboux frame by t: (3.12) On Characterization of Inextensible Flows of... 0= 107 ∂ ∂eɶ ɶ ∂tɶ ∂tɶ eɶ, tɶ = , t + eɶ, = 1 + eɶ, , ∂t ∂t ∂t ∂t ∂f ∂ ∂eɶ ∂gɶ ∂gɶ eɶ, gɶ = , gɶ + eɶ, = f 2γ + 3 + eɶ, , ∂t ∂t ∂t ∂s ∂t ∂ ɶ ∂tɶ ∂gɶ ∂gɶ 0= t , gɶ = , gɶ + tɶ, = ψ + tɶ, . ∂t ∂t ∂t ∂t 0= Considering ∂tɶ ɶ ∂gɶ ,t = , gɶ = 0 and from above statement, we obtain ∂t ∂t ∂tɶ = −eɶ +ψ gɶ , ∂t ∂f ∂gɶ = − f 2γ + 3 eɶ −ψ tɶ, ∂t ∂s where ψ = ∂tɶ , gɶ . ∂t The following theorem states the conditions on the dual geodesic curvature for the dual curve flow xɶ ( s , t ) to be inextensible. ∂xɶ = f1eɶ + f 2tɶ + f 3 gɶ is inextensible. Then, the following ∂u system of partial differential equations holds: Theorem 3.5: Suppose ∂f ∂ψ ∂γ = f 2γ + 3 + . ∂t ∂s ∂s Proof: Using (3.11), we have ∂f3 ∂ ∂gɶ ∂ = − f 2γ + eɶ −ψ tɶ ∂s ∂ t ∂s ∂s ∂ ( − f 2γ ) ∂ 2 f 3 ∂f = + 2 eɶ − f 2γ + 3 tɶ ∂s ∂s ∂s ∂ψ ɶ − t −ψ (eɶ + γ gɶ ). ∂s Therefore from dual Darboux frame ∂ ∂gɶ ∂ ∂γ ɶ = (−γ tɶ ) = − t − γ (−eɶ + ψ gɶ ). ∂s ∂ t ∂s ∂s (3.13) 108 Zehra Ekinci Thus from two equations we have (3.13). 4 Conclusion Inextensible time evolutions of curves and surfaces have an important role in computer vision, robotics and physical science. In this paper inextensible flows of dual curves according to dual darboux frame have given by considering important role of dual geometry. Since dual geometry have a significant role robotics, mechanism and dynamics, this study can be shed light on these areas. References [1] [2] [3] [4] [5] [6] [7] [8] [9] [10] [11] [12] [13] [14] [15] A. Karger and J. Novak, Space Kinematics and Lie Groups, STNL Publishers of Technical Lit., Prague, Czechoslovakia, (1978). D.Y. Know and F.C. Park, Evolution of inelastic plane curves, Appl. Math. Lett., 12(1999), 115-119. D.Y. Know, F.C. Park and D.P. Chi, Inextensible flows of curves and developable surfaces, Appl. Math. Lett., 18(2005), 1156-1162. F.M. Dimentberg, The screw calculus and its applications in mechanics, Izdat. Nauka, Moscow, USSR, (1965), English translation: AD680993, Clearing House for Federal and Scientific Technical Information. G. Chirikjian and J. Burdick, A modal approach to hyper-redundant manipulator kinematics, IEEE Trans. Robot. Autom., 10(1994), 343-354. G.S. Chirikjian, Closed-form primitives for generating volume preserving deformations, ASME J. Mechanical Design, 117(1995), 347-354. G.R. Weldkamp, On the use of dual numbers, vectors and matrices in instantaneous spatial kinematics, Mechanism and Machine Theory, II(1976), 141-156. H.H. Hacısalihoğlu, Hareket Geometrisi ve Kuaterniyonlar Teorisi, Gazi Üniversitesi Fen-Edb. Fakültesi, (1983). H. Mochimaya, E. Shimemura and H. Kobayaski, Shape control of manipulators with hyper degrees of freedom, Int. J. Robot. Res., 18(1999), 584-600. M. Desbrun and M.P. Cani-Gascuel, Active implicit surface for animation, In: Proc. Graphics Interface-Canadian Inf. Procces. Soc., (1998), 143150. M. Gage and R.S. Hamilton, The heat equation shrinking convex plane curves, J. Differential Geom., 23(1986), 69-96. M. Gage, On an area-preserving evolution equation for plane curves, Contemp. Math., 51(1986), 51-62. M. Grayson, The heat equation shrinks embedded plane curves to round points, J. Differential Geom., 26(1986), 285-314. M. Kass, A. Witkin and D. Terzopoulos, Snakes: Active contour models, In: Proc. 1 st Int. Conference on Computer Vision, (1987), 259-268. W. Blaschke, Differential Geometrie and Geometrischke Grundlagen ven Einsteins Relativitasttheorie, Dover, New York, (1945).