Coordinate system with easy linear point transformations in hyperbolic plane
Clash Royale CLAN TAG#URR8PPP
up vote
1
down vote
favorite
I want to do some analytic geometry in the hyperbolic plane. I haven't chosen a coordinate system yet since I don't know which would make the math easiest. It's a lot to ask for, but I want the description of the coordinate system (map and inverse map for these coordinates to hyperboloid model coordinates) and closed form expressions for as many of these as possible (intermediate variables is okay if it reduces the work needed to calculate these). Also avoid unsolveable integrals. The rest I'll try to fill in myself. I just don't want to be doing the math for all the different options myself when someone else has probably already done it before.
In my descriptions I treat points as interchangeable with vectors, though I don't assume operations that work in the Euclidean plane also work in the hyperbolic plane.
Geometry requires the origin be an actual point, though usually we assume it to be $O(0,0)$ or equivalent to simplify math.
Point Magnitude
Given origin point $O$ and point $A$. $|A|$ is the length of the shortest path from $O$ to $A$.
Point Negation
Given origin point $O$ and point $A$. $-A$ is $A$ rotated $180^circ$ about $O$, or $A$ scaled by $-1$ relative to $O$.
This should be easy to calculate, if it isn't just something like $(x,y)to (-x,-y)$.
Point Axis Reflection
Given origin point $O$ and point $A$. $overline A$ is $A$ reflected in the $x$ axis for $O$.
This should also be easy to calculate, if it's not just something like $(x,y)to (x,-y)$.
Point Translation
Given origin point $O$, point $A$, and translation vector $B$. The translated point $C$ is whatever $A$ would be called if $-B$ was the origin, setting the rotation so that $O$ to the new origin $-B$ looks the same as $B$ to the old origin $O$.
This translation is the basis of the new non-commutative addition and subtraction operators: $C=A+B,C-B=A$. Since the side-angle-side congruence still holds, we have an alternative choice for the geometrical interpretation:
Don't mind the straight lines. $A+Bne B+A$ because you can't have parallelograms without parallel lines. Other working geometric interpretations reduce to this or produce the same triangle.
Because this is the hyperbolic plane, translations result in rotations. This is okay.
Using this and magnitude, the equation for line length $|A-B|$ can be derived.
Point Rotation
Given origin point $O$, point $A$, and angle $theta$. The rotated point $B$ is $A$ rotated by $theta$ about the origin, satisfying $|A|=|B|,angle BOA=theta$.
Since translations also rotate, this can be derived the hard way from a series of translations.
Point Scale
Given origin point $O$, point $A$, and scale factor $c$. The scaled point $B$ is $A$ scaled by $c$ relative to the origin, or $B=cA$ in normal terms, satisfying $hat B=hat chat A,|B|=|c||A|$, using $hat x$ to denote the direction/sign of $x$. Negative scale can alternatively be calculated as $B=(-c)(-A)$.
Rectangles don't work like in Euclidean space, so it's difficult to define non-uniform scale, hence I only ask for uniform.
Using this and translation, we can use linear interpolation to get a line equation $C=(1-t)A+tB=A+t(B-A)$.
Polar coordinates seem like a pretty good candidate except for translation:
- Mapping: $(r,theta)to (kcostheta,ksintheta,z)$ using $z=cosh(r),k=sinh(r)=sqrtz^2-1$
- Inverse mapping: $(x,y,z)to (textacosh(z),textatan_2(y,x))$ using atan2
- Magnitude: $(r,theta)to r$
- Negation: $(r,theta)to (r,theta+pi)$
- Reflection: $(r,theta)to (r,-theta)$
- Translation: $(r_1,theta_1),(r_2,theta_2)totextprobably something nasty$
- Rotation: $(r,theta),alphato (r,theta+alpha)$
- Scale: $(r,theta),cto (rc,theta)text or (-rc,theta+pi)$
There's still other obscure coordinate systems that might be easier to work with, being a little more difficult for the others but having easier translation. So I'm not settling on polar yet.
linear-transformations coordinate-systems hyperbolic-geometry
add a comment |Â
up vote
1
down vote
favorite
I want to do some analytic geometry in the hyperbolic plane. I haven't chosen a coordinate system yet since I don't know which would make the math easiest. It's a lot to ask for, but I want the description of the coordinate system (map and inverse map for these coordinates to hyperboloid model coordinates) and closed form expressions for as many of these as possible (intermediate variables is okay if it reduces the work needed to calculate these). Also avoid unsolveable integrals. The rest I'll try to fill in myself. I just don't want to be doing the math for all the different options myself when someone else has probably already done it before.
In my descriptions I treat points as interchangeable with vectors, though I don't assume operations that work in the Euclidean plane also work in the hyperbolic plane.
Geometry requires the origin be an actual point, though usually we assume it to be $O(0,0)$ or equivalent to simplify math.
Point Magnitude
Given origin point $O$ and point $A$. $|A|$ is the length of the shortest path from $O$ to $A$.
Point Negation
Given origin point $O$ and point $A$. $-A$ is $A$ rotated $180^circ$ about $O$, or $A$ scaled by $-1$ relative to $O$.
This should be easy to calculate, if it isn't just something like $(x,y)to (-x,-y)$.
Point Axis Reflection
Given origin point $O$ and point $A$. $overline A$ is $A$ reflected in the $x$ axis for $O$.
This should also be easy to calculate, if it's not just something like $(x,y)to (x,-y)$.
Point Translation
Given origin point $O$, point $A$, and translation vector $B$. The translated point $C$ is whatever $A$ would be called if $-B$ was the origin, setting the rotation so that $O$ to the new origin $-B$ looks the same as $B$ to the old origin $O$.
This translation is the basis of the new non-commutative addition and subtraction operators: $C=A+B,C-B=A$. Since the side-angle-side congruence still holds, we have an alternative choice for the geometrical interpretation:
Don't mind the straight lines. $A+Bne B+A$ because you can't have parallelograms without parallel lines. Other working geometric interpretations reduce to this or produce the same triangle.
Because this is the hyperbolic plane, translations result in rotations. This is okay.
Using this and magnitude, the equation for line length $|A-B|$ can be derived.
Point Rotation
Given origin point $O$, point $A$, and angle $theta$. The rotated point $B$ is $A$ rotated by $theta$ about the origin, satisfying $|A|=|B|,angle BOA=theta$.
Since translations also rotate, this can be derived the hard way from a series of translations.
Point Scale
Given origin point $O$, point $A$, and scale factor $c$. The scaled point $B$ is $A$ scaled by $c$ relative to the origin, or $B=cA$ in normal terms, satisfying $hat B=hat chat A,|B|=|c||A|$, using $hat x$ to denote the direction/sign of $x$. Negative scale can alternatively be calculated as $B=(-c)(-A)$.
Rectangles don't work like in Euclidean space, so it's difficult to define non-uniform scale, hence I only ask for uniform.
Using this and translation, we can use linear interpolation to get a line equation $C=(1-t)A+tB=A+t(B-A)$.
Polar coordinates seem like a pretty good candidate except for translation:
- Mapping: $(r,theta)to (kcostheta,ksintheta,z)$ using $z=cosh(r),k=sinh(r)=sqrtz^2-1$
- Inverse mapping: $(x,y,z)to (textacosh(z),textatan_2(y,x))$ using atan2
- Magnitude: $(r,theta)to r$
- Negation: $(r,theta)to (r,theta+pi)$
- Reflection: $(r,theta)to (r,-theta)$
- Translation: $(r_1,theta_1),(r_2,theta_2)totextprobably something nasty$
- Rotation: $(r,theta),alphato (r,theta+alpha)$
- Scale: $(r,theta),cto (rc,theta)text or (-rc,theta+pi)$
There's still other obscure coordinate systems that might be easier to work with, being a little more difficult for the others but having easier translation. So I'm not settling on polar yet.
linear-transformations coordinate-systems hyperbolic-geometry
2
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
1
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
1
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09
add a comment |Â
up vote
1
down vote
favorite
up vote
1
down vote
favorite
I want to do some analytic geometry in the hyperbolic plane. I haven't chosen a coordinate system yet since I don't know which would make the math easiest. It's a lot to ask for, but I want the description of the coordinate system (map and inverse map for these coordinates to hyperboloid model coordinates) and closed form expressions for as many of these as possible (intermediate variables is okay if it reduces the work needed to calculate these). Also avoid unsolveable integrals. The rest I'll try to fill in myself. I just don't want to be doing the math for all the different options myself when someone else has probably already done it before.
In my descriptions I treat points as interchangeable with vectors, though I don't assume operations that work in the Euclidean plane also work in the hyperbolic plane.
Geometry requires the origin be an actual point, though usually we assume it to be $O(0,0)$ or equivalent to simplify math.
Point Magnitude
Given origin point $O$ and point $A$. $|A|$ is the length of the shortest path from $O$ to $A$.
Point Negation
Given origin point $O$ and point $A$. $-A$ is $A$ rotated $180^circ$ about $O$, or $A$ scaled by $-1$ relative to $O$.
This should be easy to calculate, if it isn't just something like $(x,y)to (-x,-y)$.
Point Axis Reflection
Given origin point $O$ and point $A$. $overline A$ is $A$ reflected in the $x$ axis for $O$.
This should also be easy to calculate, if it's not just something like $(x,y)to (x,-y)$.
Point Translation
Given origin point $O$, point $A$, and translation vector $B$. The translated point $C$ is whatever $A$ would be called if $-B$ was the origin, setting the rotation so that $O$ to the new origin $-B$ looks the same as $B$ to the old origin $O$.
This translation is the basis of the new non-commutative addition and subtraction operators: $C=A+B,C-B=A$. Since the side-angle-side congruence still holds, we have an alternative choice for the geometrical interpretation:
Don't mind the straight lines. $A+Bne B+A$ because you can't have parallelograms without parallel lines. Other working geometric interpretations reduce to this or produce the same triangle.
Because this is the hyperbolic plane, translations result in rotations. This is okay.
Using this and magnitude, the equation for line length $|A-B|$ can be derived.
Point Rotation
Given origin point $O$, point $A$, and angle $theta$. The rotated point $B$ is $A$ rotated by $theta$ about the origin, satisfying $|A|=|B|,angle BOA=theta$.
Since translations also rotate, this can be derived the hard way from a series of translations.
Point Scale
Given origin point $O$, point $A$, and scale factor $c$. The scaled point $B$ is $A$ scaled by $c$ relative to the origin, or $B=cA$ in normal terms, satisfying $hat B=hat chat A,|B|=|c||A|$, using $hat x$ to denote the direction/sign of $x$. Negative scale can alternatively be calculated as $B=(-c)(-A)$.
Rectangles don't work like in Euclidean space, so it's difficult to define non-uniform scale, hence I only ask for uniform.
Using this and translation, we can use linear interpolation to get a line equation $C=(1-t)A+tB=A+t(B-A)$.
Polar coordinates seem like a pretty good candidate except for translation:
- Mapping: $(r,theta)to (kcostheta,ksintheta,z)$ using $z=cosh(r),k=sinh(r)=sqrtz^2-1$
- Inverse mapping: $(x,y,z)to (textacosh(z),textatan_2(y,x))$ using atan2
- Magnitude: $(r,theta)to r$
- Negation: $(r,theta)to (r,theta+pi)$
- Reflection: $(r,theta)to (r,-theta)$
- Translation: $(r_1,theta_1),(r_2,theta_2)totextprobably something nasty$
- Rotation: $(r,theta),alphato (r,theta+alpha)$
- Scale: $(r,theta),cto (rc,theta)text or (-rc,theta+pi)$
There's still other obscure coordinate systems that might be easier to work with, being a little more difficult for the others but having easier translation. So I'm not settling on polar yet.
linear-transformations coordinate-systems hyperbolic-geometry
I want to do some analytic geometry in the hyperbolic plane. I haven't chosen a coordinate system yet since I don't know which would make the math easiest. It's a lot to ask for, but I want the description of the coordinate system (map and inverse map for these coordinates to hyperboloid model coordinates) and closed form expressions for as many of these as possible (intermediate variables is okay if it reduces the work needed to calculate these). Also avoid unsolveable integrals. The rest I'll try to fill in myself. I just don't want to be doing the math for all the different options myself when someone else has probably already done it before.
In my descriptions I treat points as interchangeable with vectors, though I don't assume operations that work in the Euclidean plane also work in the hyperbolic plane.
Geometry requires the origin be an actual point, though usually we assume it to be $O(0,0)$ or equivalent to simplify math.
Point Magnitude
Given origin point $O$ and point $A$. $|A|$ is the length of the shortest path from $O$ to $A$.
Point Negation
Given origin point $O$ and point $A$. $-A$ is $A$ rotated $180^circ$ about $O$, or $A$ scaled by $-1$ relative to $O$.
This should be easy to calculate, if it isn't just something like $(x,y)to (-x,-y)$.
Point Axis Reflection
Given origin point $O$ and point $A$. $overline A$ is $A$ reflected in the $x$ axis for $O$.
This should also be easy to calculate, if it's not just something like $(x,y)to (x,-y)$.
Point Translation
Given origin point $O$, point $A$, and translation vector $B$. The translated point $C$ is whatever $A$ would be called if $-B$ was the origin, setting the rotation so that $O$ to the new origin $-B$ looks the same as $B$ to the old origin $O$.
This translation is the basis of the new non-commutative addition and subtraction operators: $C=A+B,C-B=A$. Since the side-angle-side congruence still holds, we have an alternative choice for the geometrical interpretation:
Don't mind the straight lines. $A+Bne B+A$ because you can't have parallelograms without parallel lines. Other working geometric interpretations reduce to this or produce the same triangle.
Because this is the hyperbolic plane, translations result in rotations. This is okay.
Using this and magnitude, the equation for line length $|A-B|$ can be derived.
Point Rotation
Given origin point $O$, point $A$, and angle $theta$. The rotated point $B$ is $A$ rotated by $theta$ about the origin, satisfying $|A|=|B|,angle BOA=theta$.
Since translations also rotate, this can be derived the hard way from a series of translations.
Point Scale
Given origin point $O$, point $A$, and scale factor $c$. The scaled point $B$ is $A$ scaled by $c$ relative to the origin, or $B=cA$ in normal terms, satisfying $hat B=hat chat A,|B|=|c||A|$, using $hat x$ to denote the direction/sign of $x$. Negative scale can alternatively be calculated as $B=(-c)(-A)$.
Rectangles don't work like in Euclidean space, so it's difficult to define non-uniform scale, hence I only ask for uniform.
Using this and translation, we can use linear interpolation to get a line equation $C=(1-t)A+tB=A+t(B-A)$.
Polar coordinates seem like a pretty good candidate except for translation:
- Mapping: $(r,theta)to (kcostheta,ksintheta,z)$ using $z=cosh(r),k=sinh(r)=sqrtz^2-1$
- Inverse mapping: $(x,y,z)to (textacosh(z),textatan_2(y,x))$ using atan2
- Magnitude: $(r,theta)to r$
- Negation: $(r,theta)to (r,theta+pi)$
- Reflection: $(r,theta)to (r,-theta)$
- Translation: $(r_1,theta_1),(r_2,theta_2)totextprobably something nasty$
- Rotation: $(r,theta),alphato (r,theta+alpha)$
- Scale: $(r,theta),cto (rc,theta)text or (-rc,theta+pi)$
There's still other obscure coordinate systems that might be easier to work with, being a little more difficult for the others but having easier translation. So I'm not settling on polar yet.
linear-transformations coordinate-systems hyperbolic-geometry
edited Mar 5 at 6:28
asked Mar 4 at 23:21
EPICI
757
757
2
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
1
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
1
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09
add a comment |Â
2
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
1
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
1
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09
2
2
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
1
1
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
1
1
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09
add a comment |Â
1 Answer
1
active
oldest
votes
up vote
1
down vote
accepted
Here are six simple coordinate systems, using the hyperboloid model for comparison. The metric is determined by $ds^2=dx^2+dy^2-dz^2$.
Polar (circular):
$$beginalign &x=sinhvarphicostheta \ &y=sinhvarphisintheta \ &z=coshvarphi endalign$$
$$ds^2=dvarphi^2+sinh^2varphi,dtheta^2$$
Rotation by angle $alpha$ around the point $varphi=0$ : $quadthetamapstotheta+alpha$
Reflection across the line $theta=alpha$ : $quadthetamapsto2alpha-theta$
Isothermal polar:
$$beginalign &x=textcsch,rhocostheta \ &y=textcsch,rhosintheta \ &z=coth,rho endalign$$
$$ds^2=textcsch^2rho,(drho^2+dtheta^2)$$
Fermi coordinates (hypercyclic):
$$beginalign &x=sinh ucosh v \ &y=sinh v \ &z=cosh ucosh v endalign$$
$$ds^2=du^2cosh^2v+dv^2$$
Translation by distance $alpha$ along the line $v=0$ : $quad umapsto u+alpha$
(Points not on this line will move a greater distance than $alpha$, but will maintain their distance $v$ from the line.)
Reflection across the line $u=alpha$ : $quad umapsto2alpha-u$
Isothermal Fermi:
$$beginalign &x=sinh usecomega \ &y=tanomega \ &z=cosh usecomega endalign$$
$$ds^2=sec^2omega,(du^2+domega^2)$$
Horocyclic:
$$beginalign &x=Ue^-V \ &y=tfrac12U^2e^-V+sinh V \ &z=tfrac12U^2e^-V+cosh V endalign$$
$$ds^2=e^-2VdU^2+dV^2$$
"Rotation"/"Translation" by length $alpha$ along the horocycle $V=0$ : $quad Umapsto U+alpha$
(Points not on this horocycle will maintain their distance $V$ from it.)
Reflection across the line $U=alpha$ : $quad Umapsto2alpha-U$
Isothermal horocyclic (half-plane):
$$beginalign &x=(2U)/(2W) \ &y=(U^2+W^2-1)/(2W) \ &z=(U^2+W^2+1)/(2W) endalign$$
$$ds^2=tfrac1W^2(dU^2+dW^2)$$
I agree with ZenoRogue that the hyperboloid model is the easiest and most useful. Isometries of the hyperbolic plane correspond to pseudo-Euclidean isometries of the hyperboloid. Such a transformation can be represented not only by a matrix, but also by multiplication of vectors.
For example, the reflection of a point $vec p$ across a plane (in psEuc space) with normal vector $vec n$ is
$$pmapsto-n^-1pn$$
The inverse is $n^-1=frac1n^2n=frac1ncdot nn$. This points in the same direction as $n$, but has reciprocal magnitude. (Note that a psEuc vector may square to a negative number, so its inverse points in the opposite direction.)
Any isometry can be composed from reflections. Two consecutive reflections across planes intersecting at an angle $alpha$ make a rotation by angle $2alpha$. If the two planes' normals are $n_1$ and $n_2$, then the rotation is
$$pmapsto n_2^-1n_1^-1pn_1n_2=(n_1n_2)^-1p(n_1n_2)$$
add a comment |Â
1 Answer
1
active
oldest
votes
1 Answer
1
active
oldest
votes
active
oldest
votes
active
oldest
votes
up vote
1
down vote
accepted
Here are six simple coordinate systems, using the hyperboloid model for comparison. The metric is determined by $ds^2=dx^2+dy^2-dz^2$.
Polar (circular):
$$beginalign &x=sinhvarphicostheta \ &y=sinhvarphisintheta \ &z=coshvarphi endalign$$
$$ds^2=dvarphi^2+sinh^2varphi,dtheta^2$$
Rotation by angle $alpha$ around the point $varphi=0$ : $quadthetamapstotheta+alpha$
Reflection across the line $theta=alpha$ : $quadthetamapsto2alpha-theta$
Isothermal polar:
$$beginalign &x=textcsch,rhocostheta \ &y=textcsch,rhosintheta \ &z=coth,rho endalign$$
$$ds^2=textcsch^2rho,(drho^2+dtheta^2)$$
Fermi coordinates (hypercyclic):
$$beginalign &x=sinh ucosh v \ &y=sinh v \ &z=cosh ucosh v endalign$$
$$ds^2=du^2cosh^2v+dv^2$$
Translation by distance $alpha$ along the line $v=0$ : $quad umapsto u+alpha$
(Points not on this line will move a greater distance than $alpha$, but will maintain their distance $v$ from the line.)
Reflection across the line $u=alpha$ : $quad umapsto2alpha-u$
Isothermal Fermi:
$$beginalign &x=sinh usecomega \ &y=tanomega \ &z=cosh usecomega endalign$$
$$ds^2=sec^2omega,(du^2+domega^2)$$
Horocyclic:
$$beginalign &x=Ue^-V \ &y=tfrac12U^2e^-V+sinh V \ &z=tfrac12U^2e^-V+cosh V endalign$$
$$ds^2=e^-2VdU^2+dV^2$$
"Rotation"/"Translation" by length $alpha$ along the horocycle $V=0$ : $quad Umapsto U+alpha$
(Points not on this horocycle will maintain their distance $V$ from it.)
Reflection across the line $U=alpha$ : $quad Umapsto2alpha-U$
Isothermal horocyclic (half-plane):
$$beginalign &x=(2U)/(2W) \ &y=(U^2+W^2-1)/(2W) \ &z=(U^2+W^2+1)/(2W) endalign$$
$$ds^2=tfrac1W^2(dU^2+dW^2)$$
I agree with ZenoRogue that the hyperboloid model is the easiest and most useful. Isometries of the hyperbolic plane correspond to pseudo-Euclidean isometries of the hyperboloid. Such a transformation can be represented not only by a matrix, but also by multiplication of vectors.
For example, the reflection of a point $vec p$ across a plane (in psEuc space) with normal vector $vec n$ is
$$pmapsto-n^-1pn$$
The inverse is $n^-1=frac1n^2n=frac1ncdot nn$. This points in the same direction as $n$, but has reciprocal magnitude. (Note that a psEuc vector may square to a negative number, so its inverse points in the opposite direction.)
Any isometry can be composed from reflections. Two consecutive reflections across planes intersecting at an angle $alpha$ make a rotation by angle $2alpha$. If the two planes' normals are $n_1$ and $n_2$, then the rotation is
$$pmapsto n_2^-1n_1^-1pn_1n_2=(n_1n_2)^-1p(n_1n_2)$$
add a comment |Â
up vote
1
down vote
accepted
Here are six simple coordinate systems, using the hyperboloid model for comparison. The metric is determined by $ds^2=dx^2+dy^2-dz^2$.
Polar (circular):
$$beginalign &x=sinhvarphicostheta \ &y=sinhvarphisintheta \ &z=coshvarphi endalign$$
$$ds^2=dvarphi^2+sinh^2varphi,dtheta^2$$
Rotation by angle $alpha$ around the point $varphi=0$ : $quadthetamapstotheta+alpha$
Reflection across the line $theta=alpha$ : $quadthetamapsto2alpha-theta$
Isothermal polar:
$$beginalign &x=textcsch,rhocostheta \ &y=textcsch,rhosintheta \ &z=coth,rho endalign$$
$$ds^2=textcsch^2rho,(drho^2+dtheta^2)$$
Fermi coordinates (hypercyclic):
$$beginalign &x=sinh ucosh v \ &y=sinh v \ &z=cosh ucosh v endalign$$
$$ds^2=du^2cosh^2v+dv^2$$
Translation by distance $alpha$ along the line $v=0$ : $quad umapsto u+alpha$
(Points not on this line will move a greater distance than $alpha$, but will maintain their distance $v$ from the line.)
Reflection across the line $u=alpha$ : $quad umapsto2alpha-u$
Isothermal Fermi:
$$beginalign &x=sinh usecomega \ &y=tanomega \ &z=cosh usecomega endalign$$
$$ds^2=sec^2omega,(du^2+domega^2)$$
Horocyclic:
$$beginalign &x=Ue^-V \ &y=tfrac12U^2e^-V+sinh V \ &z=tfrac12U^2e^-V+cosh V endalign$$
$$ds^2=e^-2VdU^2+dV^2$$
"Rotation"/"Translation" by length $alpha$ along the horocycle $V=0$ : $quad Umapsto U+alpha$
(Points not on this horocycle will maintain their distance $V$ from it.)
Reflection across the line $U=alpha$ : $quad Umapsto2alpha-U$
Isothermal horocyclic (half-plane):
$$beginalign &x=(2U)/(2W) \ &y=(U^2+W^2-1)/(2W) \ &z=(U^2+W^2+1)/(2W) endalign$$
$$ds^2=tfrac1W^2(dU^2+dW^2)$$
I agree with ZenoRogue that the hyperboloid model is the easiest and most useful. Isometries of the hyperbolic plane correspond to pseudo-Euclidean isometries of the hyperboloid. Such a transformation can be represented not only by a matrix, but also by multiplication of vectors.
For example, the reflection of a point $vec p$ across a plane (in psEuc space) with normal vector $vec n$ is
$$pmapsto-n^-1pn$$
The inverse is $n^-1=frac1n^2n=frac1ncdot nn$. This points in the same direction as $n$, but has reciprocal magnitude. (Note that a psEuc vector may square to a negative number, so its inverse points in the opposite direction.)
Any isometry can be composed from reflections. Two consecutive reflections across planes intersecting at an angle $alpha$ make a rotation by angle $2alpha$. If the two planes' normals are $n_1$ and $n_2$, then the rotation is
$$pmapsto n_2^-1n_1^-1pn_1n_2=(n_1n_2)^-1p(n_1n_2)$$
add a comment |Â
up vote
1
down vote
accepted
up vote
1
down vote
accepted
Here are six simple coordinate systems, using the hyperboloid model for comparison. The metric is determined by $ds^2=dx^2+dy^2-dz^2$.
Polar (circular):
$$beginalign &x=sinhvarphicostheta \ &y=sinhvarphisintheta \ &z=coshvarphi endalign$$
$$ds^2=dvarphi^2+sinh^2varphi,dtheta^2$$
Rotation by angle $alpha$ around the point $varphi=0$ : $quadthetamapstotheta+alpha$
Reflection across the line $theta=alpha$ : $quadthetamapsto2alpha-theta$
Isothermal polar:
$$beginalign &x=textcsch,rhocostheta \ &y=textcsch,rhosintheta \ &z=coth,rho endalign$$
$$ds^2=textcsch^2rho,(drho^2+dtheta^2)$$
Fermi coordinates (hypercyclic):
$$beginalign &x=sinh ucosh v \ &y=sinh v \ &z=cosh ucosh v endalign$$
$$ds^2=du^2cosh^2v+dv^2$$
Translation by distance $alpha$ along the line $v=0$ : $quad umapsto u+alpha$
(Points not on this line will move a greater distance than $alpha$, but will maintain their distance $v$ from the line.)
Reflection across the line $u=alpha$ : $quad umapsto2alpha-u$
Isothermal Fermi:
$$beginalign &x=sinh usecomega \ &y=tanomega \ &z=cosh usecomega endalign$$
$$ds^2=sec^2omega,(du^2+domega^2)$$
Horocyclic:
$$beginalign &x=Ue^-V \ &y=tfrac12U^2e^-V+sinh V \ &z=tfrac12U^2e^-V+cosh V endalign$$
$$ds^2=e^-2VdU^2+dV^2$$
"Rotation"/"Translation" by length $alpha$ along the horocycle $V=0$ : $quad Umapsto U+alpha$
(Points not on this horocycle will maintain their distance $V$ from it.)
Reflection across the line $U=alpha$ : $quad Umapsto2alpha-U$
Isothermal horocyclic (half-plane):
$$beginalign &x=(2U)/(2W) \ &y=(U^2+W^2-1)/(2W) \ &z=(U^2+W^2+1)/(2W) endalign$$
$$ds^2=tfrac1W^2(dU^2+dW^2)$$
I agree with ZenoRogue that the hyperboloid model is the easiest and most useful. Isometries of the hyperbolic plane correspond to pseudo-Euclidean isometries of the hyperboloid. Such a transformation can be represented not only by a matrix, but also by multiplication of vectors.
For example, the reflection of a point $vec p$ across a plane (in psEuc space) with normal vector $vec n$ is
$$pmapsto-n^-1pn$$
The inverse is $n^-1=frac1n^2n=frac1ncdot nn$. This points in the same direction as $n$, but has reciprocal magnitude. (Note that a psEuc vector may square to a negative number, so its inverse points in the opposite direction.)
Any isometry can be composed from reflections. Two consecutive reflections across planes intersecting at an angle $alpha$ make a rotation by angle $2alpha$. If the two planes' normals are $n_1$ and $n_2$, then the rotation is
$$pmapsto n_2^-1n_1^-1pn_1n_2=(n_1n_2)^-1p(n_1n_2)$$
Here are six simple coordinate systems, using the hyperboloid model for comparison. The metric is determined by $ds^2=dx^2+dy^2-dz^2$.
Polar (circular):
$$beginalign &x=sinhvarphicostheta \ &y=sinhvarphisintheta \ &z=coshvarphi endalign$$
$$ds^2=dvarphi^2+sinh^2varphi,dtheta^2$$
Rotation by angle $alpha$ around the point $varphi=0$ : $quadthetamapstotheta+alpha$
Reflection across the line $theta=alpha$ : $quadthetamapsto2alpha-theta$
Isothermal polar:
$$beginalign &x=textcsch,rhocostheta \ &y=textcsch,rhosintheta \ &z=coth,rho endalign$$
$$ds^2=textcsch^2rho,(drho^2+dtheta^2)$$
Fermi coordinates (hypercyclic):
$$beginalign &x=sinh ucosh v \ &y=sinh v \ &z=cosh ucosh v endalign$$
$$ds^2=du^2cosh^2v+dv^2$$
Translation by distance $alpha$ along the line $v=0$ : $quad umapsto u+alpha$
(Points not on this line will move a greater distance than $alpha$, but will maintain their distance $v$ from the line.)
Reflection across the line $u=alpha$ : $quad umapsto2alpha-u$
Isothermal Fermi:
$$beginalign &x=sinh usecomega \ &y=tanomega \ &z=cosh usecomega endalign$$
$$ds^2=sec^2omega,(du^2+domega^2)$$
Horocyclic:
$$beginalign &x=Ue^-V \ &y=tfrac12U^2e^-V+sinh V \ &z=tfrac12U^2e^-V+cosh V endalign$$
$$ds^2=e^-2VdU^2+dV^2$$
"Rotation"/"Translation" by length $alpha$ along the horocycle $V=0$ : $quad Umapsto U+alpha$
(Points not on this horocycle will maintain their distance $V$ from it.)
Reflection across the line $U=alpha$ : $quad Umapsto2alpha-U$
Isothermal horocyclic (half-plane):
$$beginalign &x=(2U)/(2W) \ &y=(U^2+W^2-1)/(2W) \ &z=(U^2+W^2+1)/(2W) endalign$$
$$ds^2=tfrac1W^2(dU^2+dW^2)$$
I agree with ZenoRogue that the hyperboloid model is the easiest and most useful. Isometries of the hyperbolic plane correspond to pseudo-Euclidean isometries of the hyperboloid. Such a transformation can be represented not only by a matrix, but also by multiplication of vectors.
For example, the reflection of a point $vec p$ across a plane (in psEuc space) with normal vector $vec n$ is
$$pmapsto-n^-1pn$$
The inverse is $n^-1=frac1n^2n=frac1ncdot nn$. This points in the same direction as $n$, but has reciprocal magnitude. (Note that a psEuc vector may square to a negative number, so its inverse points in the opposite direction.)
Any isometry can be composed from reflections. Two consecutive reflections across planes intersecting at an angle $alpha$ make a rotation by angle $2alpha$. If the two planes' normals are $n_1$ and $n_2$, then the rotation is
$$pmapsto n_2^-1n_1^-1pn_1n_2=(n_1n_2)^-1p(n_1n_2)$$
answered Aug 26 at 20:44
mr_e_man
1,049422
1,049422
add a comment |Â
add a comment |Â
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
StackExchange.ready(
function ()
StackExchange.openid.initPostLogin('.new-post-login', 'https%3a%2f%2fmath.stackexchange.com%2fquestions%2f2677006%2fcoordinate-system-with-easy-linear-point-transformations-in-hyperbolic-plane%23new-answer', 'question_page');
);
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Post as a guest
Sign up or log in
StackExchange.ready(function ()
StackExchange.helpers.onClickDraftSave('#login-link');
);
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
Sign up using Google
Sign up using Facebook
Sign up using Email and Password
2
Are you familiar with the unit disk and the upper half plane models of the hyperbolic plane?
â Moishe Cohen
Mar 5 at 1:29
I concur most strongly with the implied suggestion of @MoisheCohen that you familiarize yourself with the unit disk or upper half plane representation of the hyperbolic plane.
â Lubin
Mar 5 at 4:02
1
In my experience the hyperboloid model is the easiest to work with, much easier than disk/half plane/polar. My formulas are here, I think they cover most of what you want. Some of the formulas can be simplified. I use rgpushxto0 as the translation matrix.
â Zeno Rogue
Mar 6 at 11:22
1
I didn't notice before that linear transforms can be done in the hyperboloid model as linear transforms. That's very useful. Will give it a try, thanks!
â EPICI
Mar 6 at 19:09