<<

dx 0 since we saw earlier that was equal to x . ds kx0k The unit vector T gives the direction of the . It’s useful because it says which way The unit tangent vector the path is going, but doesn’t indicate how fast and curvature the object is travelling that path. Thus, T is an Math 131 Multivariate intrinsic property of the underlying curve. We can use T to study how the curve bends, since D Joyce, Spring 2014 the bend of the curve has to do with the change in the direction of the curve. Since T is a , Summary of the arclength pararamter s. we can identify it with an angle between 0 and 2π. n The length of a path x :[a, b] → R , also called The rate of change of T, therefore, has to do with its arclength, is the of its speed the rate of change of this angle, in fact, it is the s Z b Z b  2  2 of that angle. 0 dx1 dxn kx (t)k dt = + ··· + dt. Theorem 1. Let x be a path with nonzero speed. a a dt dt Then Denote the length of the path up to t as s(t), dT 1. is orthogonal to T. that is, dt dT Z t Z t 2. k k is the angular rate of change of the di- 0 0 dt s(t) = kx k = kx (τ)k dτ. rection of T, in other words kdTk = |dθ|, where dθ a a is the differential angle associated to the differential Then by the fundamental theorem of calculus, dT. s  2  2 ds 0 dx1 dxn We saw part 1 recently when we showed it was = kx (t)k = + ··· + , 2 dt dt dt true for any unit vector u : R → R , but it’s just as true when n > 2. The other statement is also which says that the derivative of the arclength s is true of any unit vector u, and it depends on the the speed. In other words, ds is the differential of differential angle dθ being equal to the du, a vec- the arclength s. The last equation can be written tor tangent to the unit (or unit ). A in a differential form as diagram helps here. q The curvature, or bend, of a curve is suppose to ds = dx2 + ··· + dx2 . 1 n be the rate of change of the direction of the curve, so that’s how we define it. The unit tangent vector and arclength. The velocity vector, v(t) = x0(t), for a path x, points in Definition 2 (curvature). Let x be a path with x0 a direction tangent to the path at the point x(t). unit tangent vector T = 0 . The curvature κ at We can normalize it to make it a unit tangent vector kx k t is the angular rate of change of T per unit change T just by dividing it by its length: in the along the path. That is, v x0 T = = . dT kvk kx0k κ(t) = . ds 0 Of course, this is only defined when x (t) is not 0. By the , this can also be written Note that T could also be defined as kdT/dtk dx κ = . T = |ds/dt| ds

1 therefore dx T (s) = ds  −a s = √ sin √ , a2+b2 a2+b2 a s b  √ cos √ , √ a2+b2 a2+b2 a2+b2 Now differentiate this direction T with respect to s to get

dT  −a s −a s  = cos √ , sin √ , 0 ds a2+b2 a2+b2 a2+b2 a2+b2 Then, leaving out the algebra, the curvature is Figure 1: Conic

dT a κ = = . ds a2+b2 that are nearly straight have nearly 0 cur- vature, while those that up tightly have high curvature. Math 131 Home Page at http://math.clarku.edu/~djoyce/ma131/ Figure 1 shows a conic helix. Its equation in cylindrical coordinates is (r, θ, z) = (t, t, t), and in rectangular coordinates (x, y, z) = (t cos t, t sin t, t). Near the apex of the cone, (0, 0, 0), it’s curved tight with a high curvature, but as it moves away from that of the gets larger and the curva- ture decreases. It’s easy to show that a circle of radius r has curvature κ = 1/r. In fact, an alternate definition for curvature is that it is the reciprocal of the radius of the circle that best fits the curve at the point in question.

Example 3 (The helix again). From the of a helix, you can expect the curvature to be the same at every point. First let’s compute the unit tangent vector T for the helix. Since

 s s bs  x(s) = a cos √ , a sin √ , √ , a2+b2 a2+b2 a2+b2

2