8.7: Geodesic Curvature of Curves on SurfacesIn lab 3 we resolved the accleration vector of a space curve X(t) into a linear combination of the tangent vector T(t) and the principal normal vector P(t); this was in the T(t), P(t), B(t) Frenet-Serret orthonormal frame. When dealing with curves on surfaces, it is more natural to express the accleration vector in terms of a different frame. If we let U(t) denote the normal vector to the normal plane, then the vectors T(t), U(t), and N(t) form an orthonormal frame where U(t) = N(t) × T(t). We saw in the first section of lab 8 that the projection of κP(t) onto N(t) gives us the normal curvature κN(t). Here we define the projection of κP(t) onto U(t) to be the geodesic curvature, κg. We then have
X''(t) = (X'(t))' = (s'(t)T(t))' = s''(t)T(t) + s'(t)T'(t) From this equation for the acceleration vector X''(t), we can derive an explicit formula for calculating the geodesic curvature of a curve on a surface. First, take the cross-product of both sides with the velocity vector X'(t) to get:
Note that this terminology justifies our usage of the term geodesic curvature for a curve in the plane in labs 1 and 2. A curve that has geodesic curvature identically equal to zero is called a geodesic for the surface.Demonstration The demonstration shows a curve on the surface X and the expression of the tangent vector as the sum of vectors in the direction of T(t), U(t), and N(t). We can also write
Exercises
2. For a curve on the unit sphere, what can be said about κN(t)? |