WEBVTT

00:00:00.010 --> 00:00:04.090
[Music Playing]

00:00:04.090 --> 00:00:08.110
Dr. Fred Roberts: In any scenario for evacuating: New York city

00:00:08.110 --> 00:00:12.530
the Port Authority bus terminal is going to
play an absolutely essential role.

00:00:12.530 --> 00:00:17.900
And the Port Authority bus terminal is the
worlds busiest bus terminal.

00:00:17.900 --> 00:00:23.070
Its got almost of a quarter of a million passengers
pass through everyday

00:00:23.070 --> 00:00:32.350
A facility that was built in the late 1940's,
after Hurricane Sandy, the lines where so long

00:00:32.350 --> 00:00:41.590
They extended out onto the street. So they
have a major crowd management problem

00:00:41.590 --> 00:00:45.940
[Music Playing]

00:00:45.940 --> 00:00:50.160
Dr. Mubbasir Kapadia: The goals of this project
are to develop and

00:00:50.160 --> 00:00:55.320
end-to-end software solution for crowd
behavior.

00:00:55.320 --> 00:00:58.620
Andrew Dobson: The Port Authority is interested
in what may happen

00:00:58.620 --> 00:01:02.700
in case there is some sort of, natural
disaster. Having temporary

00:01:02.700 --> 00:01:06.950
construction barriers perhaps changing
completely the layout of

00:01:06.950 --> 00:01:10.950
certain floors.

00:01:10.950 --> 00:01:15.090
Dr. Mubbasir Kapadia: The first step in this
process is to use

00:01:15.090 --> 00:01:21.470
state of the art LiDAR technology to create
a virtual reconstruction

00:01:21.470 --> 00:01:28.990
which essentially gives us a photorealistic
3D virtual navigable

00:01:28.990 --> 00:01:36.020
Dr. Jie Gong: And the purpose to using cutting
edge remote sensing technologies

00:01:36.020 --> 00:01:40.070
is to capture up to date geometric information.

00:01:40.070 --> 00:01:47.320
Victor Baquero: We set up the tripod, the laser
scanning goes on top of it.

00:01:47.320 --> 00:01:51.520
Yi Yu: Than we put this target around it

00:01:51.520 --> 00:01:56.710
and put it into some overlap location with the
the next scan.

00:01:56.710 --> 00:02:00.820
Victor Baquero: So that later in the computer we can
find these targets and put the pieces together.

00:02:00.820 --> 00:02:04.880
Yi Yu: And have very high accuracy to the millimeter
level.

00:02:04.880 --> 00:02:09.300
Prarthana Raja: All these scans have a lot of
noise, a lot of people

00:02:09.300 --> 00:02:12.610
and theres also a lot overlap and we had delete
all the overlaps,

00:02:12.610 --> 00:02:14.900
clean out all the scans.

00:02:14.900 --> 00:02:19.450
Dr. Mubbasir Kapadia: The next step is to use
state of the art

00:02:19.450 --> 00:02:24.960
computer motion technology to reconstruct
the movement of the crowd

00:02:24.960 --> 00:02:30.370
in the Port Authority bus terminal
from empirical observations.

00:02:30.370 --> 00:02:35.600
Once we have these two things, we
develop computation models of how humans

00:02:35.600 --> 00:02:38.850
actually move in the Port Authority.

00:02:38.850 --> 00:02:41.880
Andrew Dobson: We want to make this very
general so that its

00:02:41.880 --> 00:02:48.340
not just for the Port Authority. The idea is
we want to create a general

00:02:48.340 --> 00:02:55.000
tool that can be used for all kinds of
public facilities.

00:02:55.000 --> 00:03:00.013
[Music Stops]