Distribution of a Robot Swarm in a Square under Gravity

Requires a Wolfram Notebook System
Interact on desktop, mobile and cloud with the free Wolfram Player or other Wolfram Language products.
This Demonstration determines the mean, variance, and covariance for a very large swarm of robots as they move inside a square workplace under the influence of gravity, pointing in the direction . The swarm is large, but the robots are comparatively small and together cover a constant area
. Under gravity, they flow like a liquid, moving to one side of the workplace to form a polygonal shape.
Contributed by: Haoran Zhao and Aaron T. Becker (January 2016)
Open content licensed under CC BY-NC-SA
Snapshots
Details
The direction of the force of gravity is determined by the angle , in
, such that the swarm can assume eight different polygonal shapes. The shapes alternate between triangles and trapezoids when
, and alternate between squares with one corner removed and trapezoids when
.
Computing the means and
, variances
and
, covariance
, and correlation
requires integration over the area containing the swarm. One way is to use an indicator function
that returns 1 if the point
is inside the region containing the swarm and 0 otherwise. The formulas are as follows, integrating over the unit square with
and
from 0 to 1.
,
,
,
,
,
.
Instead of using an indicator function, the region of integration can be changed to only include the polygon containing the swarm. As an example calculation, if the force angle is , the mean when the swarm is in the lower-left corner is
for
and
for
.
A few interesting results: the correlation is maximized when the swarm has a triangular shape, and equals . The covariance of the triangle is always
. Variance is maximized in one direction and minimized in the other when the swarm is in a rectangular position. Mean positions are maximized when
is small.
Permanent Citation