Weird phenomenon with three.js plane - math

This is the first question I've ever asked on here! Apologies in advance if I've done it wrong somehow.
I have written a program which stacks up spheres in three.js.
Each sphere starts with randomly generated (within certain bounds) x and z co-ordinates, and a y co-ordinate high above the ground plane. I casts rays from each of the sphere's vertices to see how far down it can fall before it intersects with an existing mesh.
For each sphere, I test it in 80 different random xz positions, see where it can fall the furthest, and then 'drop' it into that position.
This is intended to create bubble towers like this one:
However, I have noticed that when I make the bubble radius very small and the base dimensions of the tower large, this happens:
If I turn the recursions down from 80, this effect is less apparent. For some reason, three.js seems to think that the spheres can fall further at the corners of the base square. The origin is exactly at the center of the base square - perhaps this is relevant.
When I console log all the fall-distances I'm receiving from the raycaster, they are indeed larger the further away you get from the center of the square... but only at the 11th or 12th decimal place.
This is not so much a problem I am trying to solve (I could just round fall distances to the nearest 10th decimal place before I pick the largest one), but something I am very curious about. Does anyone know why this is happening? Has anybody come across something similar to this before?
EDIT:
I edited my code to shift everything so that the origin is no longer at the center of the base square:
So am I correct in thinking... this phenomenon is something to do with distance from the origin, rather than anything relating to the surface onto which the balls are falling?

Indeed, the pattern you are seeing is exactly because the corners and edges of the bottom of your tower are furthest from the origin where you are dropping the balls. You are creating a right triangle (see image below) in which the vertical "leg" is the line from the origin from which you are dropping the balls down to the point directly below on mesh floor (at a right angle to the floor - thus the name, right triangle). The hypotenuse is always the longest leg of a right triangle, and the futher out your rays cast from the point just below the origin, the longer the hypotenuse will be, and the more your algorithm will favor that longer distance (no matter how fractional).
Increasing the size of the tower base would exaggerate this effect as the hypotenuse measurements can now grow even larger. Reducing the size of the balls would also favor the pattern you are seeing, as now each ball is not taking up as much space, and so the distant measurments to the corners won't fill in as quickly as they would with larger balls so that now more balls will congregate at the edges before filling in the rest of the space.
Moving your dropping origin to one side or another creates longer distances (hypotenuses) to the opposites sides and corners, so that the balls will fill in those distant locations first.
The reason you see less of an effect when you reduce the sample size from 80 to say, 20, is that there are simply fewer chances to detect these more distant locations to which the balls could fall (an odds game).
A right triangle:
A back-of-the-napkin sketch:

Related

How to find the sides of a rectangle if you know the sides of a quadrilateral inside the rectangle?

I'm working on an application that uses a accelerometer to measure the sides of a room, I know it will not be exact measurements but it's fine.
In reality I would like the program to be able to calculate the sides of any room shape not only rectangles and squares (and more than 4 corners), but I'm starting with something more simple (rectangle shaped rooms).
My problem is not with the accelerometer but more with the math aspect of the code. Because I measured the room by placing the phone on a wall and then going to the connected wall, I will get the measurements of a quadrilateral inside the rectangle. From there, if it's possible, I will get the measurements of the sides of the rectangle, but I don't really know how.
What I've tried so far:
Divided the quadrilateral inside the rectangle in half, to make 2 triangles. Then I calculated the diagonal using the Pythagoras theorem. Then I used the law of Cosines to calculate one of the angles, and did the same again to find another. Then found the 3rd angle using the 2 other angles (c=a+b-180). I did this for both triangles.
I don't know if this is the right approach and if I have missed something simple, or if I simply don't have enough information to solve for the sides of the rectangle. I have looked into some geometry and trigonometry math online and haven't find anything that gives me a solution. But like I said, maybe I missed something simple.
Any push in the right direction would be helpful.
The rectangle and the quadrilateral
The problem lacks a unique solution. Imagine placing a pair of calipers around the quadrilateral. You'll be able to rotate the calipers around it, and at each angle the calipers will be able to close to a different width. Each of those widths is a different possible room dimension.
You'll also never get an accurate position measurement using the inertial sensors in a phone to begin with. The accels and gyros aren't even close to accurate enough. GPS is, but only outdoors away from structures that cause multipathing artifacts. Quick and sloppy with a tape measure will win every time.

Resize image using real world measurements

I'm working on a floor design app where the user can import a floor texture and the app will place the texture on to a room image.
I've managed to transform the perspective of the floor image so that it matches the room image - thanks to this answer, but I'm now stuck on scaling the floor image to match the room image dimensions.
I know the real dimensions of the wooden floor (177mm x 1220mm per plank), I know the height of an object in the room image (height of white tile near sink is 240mm) and I know the distance between the camera and the white tile (roughly 2500mm). The room image size is 2592x1936, the floor image size is 1430x1220.
The room image was taken with from an iPad air camera to which I can't seem to find any info regarding the focal length and sensor size, the nearest I could find was a 3.3 focal length with 3.6mm sensor height (this may be where I'm going wrong).
I tried using this equation
The numbers I plugged in to the equation,
2662 = (3.3 240 x 1936) / (160 x 3.6)
I then tried to work out the object height for a wooden plank in the floor image,
(3.3 x 1220 x 1936) / (2662 x 3.6) = 813 px
I then divided the image height by the object height to get a ratio = 2.38.
This image is with a 2.38 ratio applied to the floor image which isn't quite right.
I know I'm going wrong somewhere or going the complete wrong way about it, hope somebody can point me in the right direction.
Thanks
I'd extend the lines of the tile till they touch the edge where the back wall meets the floor. Using this technique you can transfer a length from the wall plane to an equal length in the floor plane. So at that point, all you have to do is match lengths along a single line, namely the lengths between planks and the lengths between your transferred points. But you have to do this in a projectively consistent fashion. The most versatile tool for projective measurements is the cross ratio. An application very similar to what you have here is described in How to calculate true lengths from perspective projection on Math SE. If your vanishing point on that line where the walls meet is indeed at infinity (which appears to be approximately the case in your setup), you can get by with some simpler computations, but unless you can guarantee that this will always be the case, I'd not rely on that.
The above will help you adjust the scale in one direction only. The direction perpendicular to that is still open, though. In your exaple that would be the depth direction, the direction away from the camera. Do you have any reference points for that direction? It looks to me as though you might be able to use one complete tile on the left wall, before the window starts. But depending on how the corner between the two walls is tiled, that might be slightly off.
To illustrate these ideas, look at the picture above. Since the red lines appear almost horizontal, seeing the effects of perspective there is pretty hard. Therefore I'll do the other direction. Suppose the tile in the corner is indeed the same visible size as all the other tiles on the wall. So you know the real world distance between A1 and B1. You project along the blue vertical lines (vertical in the real world, not neccessarily the image) down to A2 and B2 which is where the left wall plane meets the floor plane.
Why do they meet there? Well, the lines A1,A2 is where the left all meets the back wall. The line A2,A3 is where the back wall meets the floor. Both of these plane intersections are actually visible at least in part, which made drawing the lines possible. So at A2 all three planes meet, and connecting that to the far point F gives the third edge, where the left wall meets the floor.
Since the segments A1,B1 and A2,B2 are just vertically transported versions of one another in the real world, they have equals length. That transportation was along the left wall in the vertical direction. Now transport them again, this time in the floor plane and in the left-right direction. You do so using the red lines, which are either parallel or meet at a point (which is pretty far away in this example). These red lines A2,A3 and B2,B3 are parallel in the real world, and their distance is still the edge length of that tile.
Now start measuring something, e.g. distance between C and D. To do that, compute the cross ratio (F,A3;B3,C) which expresses the distance from A3 to C, expressed in multiples of the distance from A3 to B3, and using F as the point at infinity. Do the same for D, and then the difference will be the length from C to D, expressed in multiples of the distance from A3 to B3. So the distance between C and D is 4.42 tile edge lengths in this example. Scale your image to fit this figure.

Finding the normal between a cylinder and a triangle

I have a pretty rudimentary physics engine in the game I'm working on, between moving, cylindrical characters, and static meshes made of triangles. The intended behavior is for characters to slide across surfaces, and in most cases, it works fine. But the engine doesn't discriminate between a head-on collision and a glancing collision.
I'm not entirely sure what information I could give that would be helpful. I'm looking for a mathematical solution, at any rate, a method to determine the 'angle of contact' between an arbitrary cylinder and triangle. My instincts, or whatever, tell me that I need to find the point of contact between the triangle and the cylinder, then determine whether that point is within the triangle (Using the triangle's regular normal) or along one of its edges (Using the angle between the point of contact and some point on the cylinder, I'm not sure which.), but I'm sure there's a better solution.
As requested, here's a couple of examples. In this first image, a cylinder travels downwards towards a triangle (In this example, the triangle is vertical, simplified to a line.) I project the velocity vector onto the plane of the triangle, using the formula Vf = V - N * (dot(V,N)). This is the intended behavior for this type of collision.
In this image, the cylinder's axis is parallel with the normal of the triangle. Under the current implementation, Vf is still determined using the triangle's natural normal, which would cause the cylinder to begin moving vertically. Under intended behavior, N would be perpendicular to the colliding edge of the triangle.
But these are just the two extremes of collision. There are going to be a bunch of in-betweens, so I need a more arbitrary solution.
This is my attempt at a more 3D example. I apologize for the poor perspective. The bottom-most vertex in this triangle is closer to the 'camera'. The point of collision between the cylinder and the triangle is marked by the red X. Under intended behavior, if the cylinder was moving directly away from the camera, it would slide to the left, along the length of the triangle's edge. No vertical movement would be imparted, as the point of contact is along the cylinder's, uh, tube section, rather than the caps.
Under current behavior, the triangle's normal is used. The cylinder would be pushed upwards, as though sliding across the face of the triangle, while doing little to prevent movement into the triangle.
I understand that this is a difficult request, so I appreciate the suggestions made to help refine my question.
What you're looking for is probably an edge collision detector. In rigid body collision systems, there are usually two types of collisions: surface collisions (for colliding with things that have a regular surface normal, where the reaction normal can be computed easily, as you pointed out, by processing A velocity vs B surface normal), and edge collisions (where the A body hits the edge of B body, be it box, triangle or anything else). In this case, the matter is more complicated, because, obviously, edge is not a surface, and thus you can't calculate it's normal at all. Usually, it's approximated one way or another - you can for example assume that, for triangle mesh, the edge normal is the average between normals of the two edge triangle's. There are also other methods to deal with it, some discussed here:
https://code.google.com/p/bullet/downloads/detail?name=CEDEC2011_ErwinCoumans.pdf&can=2&q=
Usually, there's an edge processing threshold value, if a collision occurred in the radius of this value, it's considered an edge collision, and processed differently.
See the examples here:
http://www.wildbunny.co.uk/blog/2012/10/31/2d-polygonal-collision-detection-and-internal-edges/
Googling "internal edge collision" and learning about rigid body collisions/dynamics in general will help you understand and solve this problem by yourself.

Calculating 2D angles for 3D objects in perspective

Imagine a photo, with the face of a building marked out.
Its given that the face of the building is a rectangle, with 90 degree corners. However, because its a photo, perspective will be involved and the parallel edges of the face will converge on the horizon.
With such a rectangle, how do you calculate the angle in 2D of the vectors of the edges of a face that is at right angles to it?
In the image below, the blue is the face marked on the photo, and I'm wondering how to calculate the 2D vector of the red lines of the other face:
example http://img689.imageshack.us/img689/2060/leslievillestarbuckscor.jpg
So if you ignore the picture for a moment, and concentrate on the lines, is there enough information in one of the face outlines - the interior angles and such - to know the path of the face on the other side of the corner? What would the formula be?
We know that both are rectangles - that is that each corner is a right angle - and that they are at right angles to each other. So how do you determine the vector of the second face using only knowledge of the position of the first?
It's quite easy, you should use basic 2 point perspective rules.
First of all you need 2 vanishing points, one to the left and one to the right of your object. They'll both stay on the same horizon line.
alt text http://img62.imageshack.us/img62/9669/perspectiveh.png
After having placed the horizon (that chooses the sight heigh) and the vanishing points (the positions of the points will change field of view) you can easily calculate where your lines go (of course you need to be able to calculate the line that crosses two points: i think you can do it)
Honestly, what I'd do is a Hough Transform on the image and determine a way to identify the red lines from the image. To find the red lines, I'd find any lines in the transform that touch your blue ones. The good thing about the transform is that you get angle information for free.
Since you know that you're looking at lines, you could also do a Radon Transform and look for peaks at particular angles; it's essentially the same thing.
Matlab has some nice functionality for this kind of work.

circles and triangles problem

I have an interesting problem here I've been trying to solve for the last little while:
I have 3 circles on a 2D xy plane, each with the same known radius. I know the coordinates of each of the three centers (they are arbitrary and can be anywhere).
What is the largest triangle that can be drawn such that each vertex of the triangle sits on a separate circle, what are the coordinates of those verticies?
I've been looking at this problem for hours and asked a bunch of people but so far only one person has been able to suggest a plausible solution (though I have no way of proving it).
The solution that we have come up with involves first creating a triangle about the three circle centers. Next we look at each circle individually and calculate the equation of a line that passes through the circle's center and is perpendicular to the opposite edge. We then calculate two intersection points of the circle. This is then done for the next two circles with a result of 6 points. We iterate over the 8 possible 3 point triangles that these 6 points create (the restriction is that each point of the big triangle must be on a separate circle) and find the maximum size.
The results look reasonable (at least when drawn out on paper) and it passes the special case of when the centers of the circles all fall on a straight line (gives a known largest triangle). Unfortunate i have no way of proving this is correct or not.
I'm wondering if anyone has encountered a problem similar to this and if so, how did you solve it?
Note: I understand that this is mostly a math question and not programming, however it is going to be implemented in code and it must be optimized to run very fast and efficient. In fact, I already have the above solution in code and tested to be working, if you would like to take a look, please let me know, i chose not to post it because its all in vector form and pretty much impossible to figure out exactly what is going on (because it's been condensed to be more efficient).
Lastly, yes this is for school work, though it is NOT a homework question/assignment/project. It's part of my graduate thesis (abet a very very small part, but still technically is part of it).
Thanks for your help.
Edit: Heres a new algorithm that i came up with a little while ago.
Starting at a circle's centre, draw a line to the other two centres. Calculate the line that bisects the angle created and calculate the intersections between the circle and the line that passes through the centre of your circle. You will get 2 results. Repeat this for the other two circles to get a total of 6 points. Iterate over these 6 points and get 8 possible solutions. Find the maximum of the 8 solutions.
This algorithm will deal with the collinear case if you draw your lines in one "direction" about the three points.
From the few random trials i have attempted using CAD software to figure out the geometries for me, this method seems to outperform all other methods previously stated However, it has already been proven to not be an optimal solution by one of Victor's counter examples.
I'll code this up tomorrow, for some reason I've lost remote access to my university computer and most things are on it.
I've taken the liberty of submitting a second answer, because my original answer referred to an online app that people could play with to get insight. The answer here is more a geometric argument.
The following diagram illuminates, I hope, what is going on. Much of this was inspired by #Federico Ramponi's observation that the largest triangle is characterized by the tangent at each vertex being parallel to the opposite side.
(source: brainjam.ca)
The picture was produced using a trial version of the excellent desktop program Geometry Expressions. The diagram shows the three circles centered at points A,E, and C. They have equal radii, but the picture doesn't really depend on the radii being equal, so the solution generalizes to circles of different radii. The lines MN, NO, and OM are tangent to the circles, and touch the circles at the points I,H, and G respectively. The latter points form the inner triangle IHG which is the triangle whose size we want to maximize.
There is also an exterior triangle MNO which is homethetic to the interior triangle, meaning that its sides are parallel to that of IHG.
#Federico observed that IHG has maximal area because moving any of its vertices along the corresponding circle will result an a triangle that has the same base but less height, therefore less area. To put it in slightly more technical terms, if the triangle is parameterized by angles t1,t2,t3 on the three circles (as pointed out by #Charles Stewart, and as used in my steepest descent canvas app), then the gradient of the area w.r.t to (t1,t2,t3) is (0,0,0), and the area is extremal (maximal in the diagram).
So how is this diagram computed? I'll admit in advance that I don't quite have the full story, but here's a start. Given the three circles, select a point M. Draw tangents to the circles centered at E and C, and designate the tangent points as G and I. Draw a tangent OHN to the circle centered at A that is parallel to GI. These are fairly straightforward operations both algebraically and geometrically.
But we aren't finished. So far we only have the condition that OHN is parallel to GI. We have no guarantee that MGO is parallel to IH or that MIN is parallel to GH. So we have to go back and refine M. In an interactive geometry program it's no big deal to set this up and then move M until the latter parallel conditions are met (by eyeballs, anyways). Geometry Expressions created the diagram, but I used a bit of a cheat to get it to do so, because its constraint solver was apparently not powerful enough to do the job. The algebraic expressions for G, I, and H are reasonably straightforward, so it should be possible to solve for M based on the fact that MIHG is a parallelogram, either explicitly or numerically.
I should point out that in general if you follow the construction starting from M, you have two choices of tangent for each circle, and therefore eight possible solutions. As in the other attempted answers to the question, unless you have a good heuristic to help you choose in advance which of the tangents to compute, you should probably compute all eight possible triangles and find the one with maximum area. The other seven will be extremal in the sense of being minimal area or saddle points.
That's it. This answer is not quite complete in that it leaves the final computation of M somewhat open ended. But it's reduced to either a 2D search space or the solution of an ornery but not humongous equation.
Finally, I have to disagree with #Federico's conclusion that this confirms that the solution proposed by the OP is optimal. It's true that if you draw perpendiculars from the circle centers to the opposite edge of the inner triangle, those perpendiculars intersect the circle to give you the triangle vertex. E.g. H lies on the line through A perpendicular to GI), but this is not the same as in the original proposed solution (which was to take the line through A and perpendicular to EC - in general EC is not parallel to GI).
I've created an HTML5 canvas app that may be useful for people to play with. It's pretty basic (and the code is not beautiful), but it lets you move three circles of equal radius, and then calculates a maximal triangle using gradient/steepest descent. You can also save bitmaps of the diagram. The diagram also shows the triangle whose vertices are the circle centers, and one of the altitudes. Edit1: the "altitude" is really just a line segment through one of the circle centers and perpendicular to the opposite edge of the triangle joining the centers. It's there because some of the suggested constructions use it. Edit2: the steepest descent method sometimes gets stuck in a local maximum. You can get out of that maximum by moving a circle until the black triangle flips and then bringing the circle back to its original position. Working on how to find the global maximum.
This won't work in IE because it doesn't support canvas, but most other "modern" browsers should work.
I did this partially because I found some of the arguments on this page questionable, and partially because I've never programmed a steepest descent and wanted to see how that worked. Anyways, I hope this helps, and I hope to weigh in with some more comments later.
Edit: I've looked at the geometry a little more and have written up my findings in a separate answer.
Let A, B, C be the vertexes of your triangle, and suppose they are placed as in your solution.
Notice that the key property of your construction is that each of the vertexes lies on a tangent to its circle which is parallel to the opposite side of the triangle. Obviously, the circle itself lies entirely on one side of the tangent, and in the optimal solution each tangent leaves its circle on the same side as the other vertexes.
Consider AB as the "base" of the triangle, and let C float in its circle. If you move C to another position C' within the circle, you will obtain another triangle ABC' with the same base but a smaller height, hence also with a smaller area:
figure 1 http://control.ee.ethz.ch/~ramponif/stuff/circles1.png
For the same reason, you can easily see that any position of the vertexes that doesn't follow your construction cannot be optimal. Suppose, for instance, that each one of the vertexes A', B', C' does not lie on a tangent parallel to the side connecting the other two.
Then, constructing the tangent to the circle that contains (say) C', which is parallel to A'B' and leaves the circle on the same side as A'B', and moving C' to the point of tangency C, it is always possible to construct a triangle A'B'C which has the same base, but a greater height, hence also a greater area:
figure 2 http://control.ee.ethz.ch/~ramponif/stuff/circles2.png
Since any triangle that does not follow your construction cannot be optimal, I do believe that your construction is optimal. In the case when the centers of the circles are aligned I'm a bit confused, but I guess that it is possible to prove optimality along the same lines.
I believe this is a convex optimization problem (no it's not, see below), and hence can be solved efficiently using well known methods.
You essentially want to solve the problem:
maximize: area(v1,v2,v3) ~ |cross((v2-v1), (v3-v1))|
such that: v1 in C1, v2 in C2, v3 in C3 (i.e., v_i-c_i)^2 - r_i^2 <= 0)
Each of the constraints are convex, and the area function is convex as well. Now, I don't know if there is a more efficient formulation, but you can at least use an interior point method with derivatives since the derivative of the area with respect to each vertex position can be worked out analytically (I have it written down somewhere...).
Edit: grad(area(v1,v2,v3))(v_i) = rot90(vec(vj,vk)), where vec(a,b) is making a 2D vector starting at a and ending at b, and rot90 means a positive orientation rotation by 90 degrees, assuming (vi,vj,vk) was positively oriented.
Edit 2: The problem is not convex, as should be obvious considering the collinear case; two degenerate solutions is a sure sign of non-convexity. However, the configuration starting at the circle centers should be in the globally optimal local maximum.
Not optimal, works well when all three are not colinear:
I don't have a proof (and therefore don't know if it's guaranteed to be biggest). Maybe I'll work on one. But:
We have three circles with radius R with positions (from center) P0, P1, and P2. We wish to find the vertices of a triangle such that the area of the triangle is maximum, and the vertices lie on any point of the circles edges.
Find the center of all the circles and call that C. Then C = (P0 + P1 + P2) / 3. Then we find the point on each circle farthest from C.
Find vectors V0, V1, and V2, where Vi = Pi - C. Then find points Q0, Q1, and Q2, where Qi = norm(Vi) * R + Pi. Where norm indicates normalization of a vector, norm(V) = V / |V|.
Q0, Q1, and Q2 are the vertices of the triangle. I assume this is optimal because this is the farthest the vertices could be from each other. (I think.)
My first thought is that you should be able to find an analytic solution.
Then the equations of the circles are:
(x1-h1)^2 + (y1-k1)^2 = r^2
(x2-h2)^2 + (y2-k2)^2 = r^2
(x3-h3)^2 + (y3-k3)^2 = r^2
The vertices of your triangle are (x1, y1), (x2, y2), and (x3, y3). The side lengths of your triangle are
A = sqrt((x1-x2)^2 + (y1-y2)^2)
B = sqrt((x1-x3)^2 + (y1-y3)^2)
C = sqrt((x2-x3)^2 + (y2-y3)^2)
So the area of the triangle is (using Heron's formula)
S = (A+B+C)/2
area = sqrt(S(S-A)(S-B)(S-C))
So area is a function of 6 variables.
At this point I realize this is not a fruitful line of reasoning. This is more like something I'd drop into a simulated annealing system.
So my second thought is to choose the point on circle with centre A as follows: Construct line BC joining the centres of the other two circles, then construct the line AD that is perpendicular to BC and passes through A. One vertex of the triangle is the intersection of AD and circle with centre A. Likewise for the other vertices. I can't prove this but I think it gives different results than the simple "furthest from the centre of all the circles" method, and for some reason it feels better to me. I know, not very mathematical, but then I'm a programmer.
Let's assume the center of the circles to be C0,C1 and C2; and the radius R.
Since the area of a triangle is .5*base*height, let's first find the maximum base that can be constructed with the circles.
Base = Max {(|C0-C1|+2R),(|C1-C2|+2R,(|C2-C0|+2R}
Once the base length is determined between 2 circles, then we can find the farthest perpendicular point from the base line to the third circle. (product of the their slopes is -1)
For special cases such as circles aligned in a single line, we need to perform additional checks at the time of determining the base line.
It appears that finding the largest Apollonius circle for the three circles and then inscribing an equilateral triangle in that circle would be a solution. Proof left as an exercise ;).
EDIT
This method has issues for collinear circles like other solutions here, too and doesn't work.
Some initial thoughts.
Definition Call the sought-after triangle, the maximal triangle. Note that this might not be unique: if the circles all have the same centre, then there are infinitely many maximal triangles obtained by rotation around the center, and if the centres are colinear, then there will be two maximal triangles, each a mirror image of the other.
Definition Call the triangle (possibly, degenerately, either a point or a line) whose vertices are the centres of the circles the interior triangle.
Observation The solution can be expressed as three angles, indicating where on the circumference of each circle the triangle is to be found.
Observation Given two exterior vertices, we can determine a third vertex that gives the maximal area: draw the altitude of the triangle between the two exterior vertices and the centre of the other circle. This line intersects the circumference in two places; the further away point is the maximising choice of third vertex. (Fixed incorrect algorithm, Federico's argument can be adapted to show correctness of this observation)
Consequence The problem is reduced to from a problem in three angles to one in two.
Conjecture Imagine the diagram is a pinboard, with three pins at the three centres of the circles. Imagine also a closed loop of string of length equal to the perimiter of the interior triangle, plus the radius of a circle, and we place this loop around the pins. Take an imaginary pen and imaginarily draw the looping figure where the loop is always tight. I conjecture that the points of the maximal triangle will all lie on this looping figure, and that in the case where the interior triangle is not degenerate, the vertices of the maximal triangle will be the three points where the looping figure intersects one of the circle circumferences. Many counterexamples
More to follow when I can spare time to think about it.
This is just a thought, no proof or math to go along with the construction just yet. It requires that the circle centers not be colinear if the radii are the same for each circle. This restriction can be relaxed if the radii are different.
Construction:
(1) Construct a triangle such that each side of the triangle is tangent to two circles, and therefore, each circle has a tangent point on two sides of the triangle.
(2) Draw the chord between these two tangent points on each circle
(3) Find the point on the boundary of the circle on the extended ray starting at the circle's center through the midpoint of the chord. There should be one such point on each of the three circles.
(4) Connect them three points of (3) to fom a triangle.
At that point I don't know if it's the largest such triangle, but if you're looking for something approximate, this might be it.
Later: You might be able to find an approximate answer for the degenerate case by perturbing the "middle" circle slightly in a direction perpendicular to the line connecting the three circles.

Resources