9

I need to know how to plot and animate a circle "rolling" along this pretty complicated 3D parametric curve.

r[t_] := {4 Sin[t] Cos[4 t], 3 Sin[t] Sin[4 t], 3 Sin[t]};
ParametricPlot3D[r[t], {t, 0, 2 \[Pi]}]

I really have no clue where to start. I can animate basic things like drawing on a curve but not something like this.

I have the unit vectors but will refrain from posting here since they are quite lengthy. T(t) is the tangent unit vector and N(t) is the normal unit vector.

0 < u < 2pi

Find an equation for the circle C_u of radius 1 that's centered at r(u) + N(u) and contained in the unit tangent vector and unit normal vector planes.

Plot the circles, u = 1/2 and u = pi with the curve.

Animate the plots of all the circles C_u, 0 < u < 2pi, with the curve. It should appear to be wheel rolling along the curve in the normal position. EDIT: The circle must be oriented "flat" (parallel to the x-plane I believe) the entire time. The first reply is really good, it just needs to be oriented flat and inside the curve at all times.

This is baffling

Thanks

EDIT: The editor won't let me put a greeting at the beginning. Also if you need me to post the unit vectors I will gladly.

EDIT 2: Here is my attempt of trying to plot just the circle:

Manipulate[ParametricPlot3D[{Circle[{t, r[t]}, 1]}, PlotRange -> Automatic], {t, 0, 2 \[Pi]}]

EDIT 3: The circle must be oriented "flat" (parallel to the x-plane I believe) the entire time. Sorry I forgot to mention this

naanman
  • 185
  • 1
  • 6

1 Answers1

21

EDIT

As OP wishes (and as Rahul correctly points out) my original answer puts unit circle in TB plane (my error as labels suggest) and what is desired is TN plane.

pp = ParametricPlot3D[r[t], {t, 0, 2 \[Pi]}]
fs = FrenetSerretSystem[r[t], t];
tan[s_] := fs[[2, 1]] /. t -> s
nrm[s_] := fs[[2, 2]] /. t -> s
cir[u_] := 
 Table[r[u] + Cos[j] tan[u] + (Sin[j] + 1) nrm[u], {j, 0, 2 Pi, 0.1}]

Export gif of:

Table[
 Show[
  Graphics3D[{Yellow, PointSize[0.04], Point[pnts[v]], 
    Thickness[0.02], Red, Line[cir[v]]}], pp, 
  PlotRange -> {{-5, 5}, {-5, 5}, {-5, 5}}, Boxed -> False, 
  Background -> Black], {v, 0, 2 Pi, 0.1}]

enter image description here

Original Answer(with variable name change)

It is relatively straightforward to slide a circle, e.g.

pp = ParametricPlot3D[r[t], {t, 0, 2 π}];
fs = FrenetSerretSystem[r[t], t];
tan[s_] := fs[[2, 1]] /. t -> s
bnrm[s_] := fs[[2, 3]] /. t -> s
cir[u_] := 
 Table[r[u] + Cos[j] tan[u] + (Sin[j] + 1) bnrm[u], {j, 0, 2 Pi, 0.1}]

So,

Manipulate[
 Show[Graphics3D[{Red, Thickness[0.02], Line[cir[v]]}], pp, 
  PlotRange -> {{-4, 4}, {-4, 4}, {-5, 5}}, Background -> Black, 
  Boxed -> False], {v, 0, 2 Pi}]

enter image description here

To simulate rolliing (the arc length expression was derived from ArcLength function:

al[u_] := 
 NIntegrate[\[Sqrt](9 Cos[t]^2 + (12 Cos[4 t] Sin[t] + 
       3 Cos[t] Sin[4 t])^2 + (4 Cos[t] Cos[4 t] - 
       16 Sin[t] Sin[4 t])^2), {t, 0, u}]
p[t_] := r[t] + Sin[al[t]] tan[t] + (1 + Cos[al[t]]) bnrm[t]

Visualizing:

Manipulate[
 Show[
  Graphics3D[{Yellow, PointSize[0.04], Point[p[v]], Thickness[0.02], 
    Red, Line[cir[v]]}], pp, PlotRange -> {{-5, 5}, {-5, 5}, {-5, 5}},
   Boxed -> False, Background -> Black], {v, 0, 2 Pi}]

enter image description here

ubpdqn
  • 60,617
  • 3
  • 59
  • 148
  • There are sections of your first GIF where the red circle disappears behind a static black square. I think the image might be a little bit corrupt (missing data on animation layers) because there's no reason for the actual function to do that. (Right?) – wizzwizz4 Oct 25 '16 at 08:35
  • @wizzwizz4 the gif is a screen capture. Thank you for your comment. If it is not a defect related to capture then I am sure better quality and better approaches are feasible. In reflection, if you are referring to 'cut off' this relates to plot range. I was too time poor to edit. Feel free to improve as you see fit.:) – ubpdqn Oct 25 '16 at 08:40
  • I can see it at ~4.9 and ~1.5 on the first image. Unfortunately I don't have Mathematica and so can't re-record it myself. As long as it's not a problem with the function and just the image it doesn't really affect the answer at all. – wizzwizz4 Oct 25 '16 at 08:43
  • @wizzwizz4 I appreciate you pointing it out. I think it is the range issue. You can see I extended the plot range in the second gif which I think has no such defects. There can be other ways to achieve the goal and you (when you have access to MMA) or others can do them. There are other ways to make my answer look nicer. I just wanted to propose a start. :) – ubpdqn Oct 25 '16 at 08:48
  • @naanman you could use the tangent and binormal unit vector as the basis for circle. Good luck with achieving your aim. :) – ubpdqn Oct 26 '16 at 00:54
  • @ubpdqn This is excellent! Except the circle needs to be constantly flat (parallel to the x-plane) I'll try to figure it out on my own but if you can show me how to do this it would very much appreciated. Thanks! – naanman Oct 26 '16 at 01:07
  • @ubpdqn I'm kinda struggling to figure out how to rotate the circle... can you show me how? What in cir[u_] controls the orientation? reversing tan and nrm just puts it on the other side of the curve, and reversing sin and cos rotates it on it's y-axis – naanman Oct 26 '16 at 01:24
  • @naanman as the parameter t increases the circle must rotate by amount equal to arc length for rolling (i.e if at time t=x the arc length is 2$\pi$ then the point on the unit circle would have completed one revolution.. In the case of unit circle (r=1) you don't need to "scale" angle. Play around and see effect. Good luck. Am a bit time poor. :) – ubpdqn Oct 26 '16 at 01:30
  • I think the normal should be fs[[2, 2]]. According to the documentation fs[[2, 3]] is the binormal. –  Oct 26 '16 at 02:04
  • @Rahul I agree from TNB basis I have put circle in TB plane and what OP desires is TN... if I get time I will correct. Thank you for pointing it out. – ubpdqn Oct 26 '16 at 02:08
  • @ubpdqn you are a wonderful person. thanks so much – naanman Oct 28 '16 at 01:30
  • @Rahul thank you for pointing that out- i was changing so many values but it did not result in what I wanted. thanks – naanman Oct 28 '16 at 01:31
  • @naanman thank you for the kind words. I too thank Rahul for his instructive and corrective comment. :) – ubpdqn Oct 28 '16 at 03:07