00:01
(Smooth techno music)
00:08
[Narrator] Drones are here to stay.
00:10
They're not going anywhere.
00:12
In the near future, we will be seeing a lot more drones
00:14
simply flying for diverse reasons, and we need
00:17
technology that can control and detect them,
00:20
be it for package delivery, for public security.
00:24
It could be like you set up a patrol drone
00:27
that's inspecting a park.
00:29
Or use of drones for bad reasons.
00:31
For drug smuggling, for carrying explosives
00:35
or radioactive material, and, of course, espionage.
00:39
Espionage not necessarily just via video and photo,
00:43
but also with Wi-Fi interceptors.
00:45
Drones can actually get the signal of a Wi-Fi router
00:49
and phish documents or other sensitive data
00:52
within your organization.
00:55
I'm the regional sales manager for the Americas for Dedrone.
00:59
Dedrone is a drone detection technology company.
01:03
We're a German startup now based in San Francisco
01:07
and we developed the DroneTracker,
01:09
which is a key infrastructure for security
01:13
and detection of drones.
01:15
We have correctional facilities; we have border control,
01:18
data centers, stadiums, and arenas.
01:21
We also have public spaces like parks or events
01:25
in public spaces, amphitheaters, et cetera,
01:28
that require technology that tracks and detects
01:31
drone activity in the lower airspace.
01:34
The DroneTracker is a sensor packet comprised of
01:38
a video sensor, two microphone or acoustic sensors
01:41
for basically all acoustic and ultrasonic.
01:45
We have an infrared sensor.
01:47
There's also a wireless receiver
01:49
or wireless sensor on this other arm.
01:52
Dedrone developed something we call DroneDNA.
01:55
So it's the very particular things that are inherent
01:59
to drones in terms of the acoustic ranges
02:02
that the motors will omit sound in.
02:05
The different radio frequencies that they
02:07
intercommunicate with the RCs,
02:09
and also the actual shape and form.
02:11
So it will basically identify by comparing
02:14
within our libraries the different
02:17
visual aspects of a drone.
02:19
A lot of video analytics takes place there.
02:22
At the end, everything is weighed in
02:24
within our algorithm to deliver a positive alert.
02:29
The objective is to distinguish definitely from a drone
02:32
and anything else that is not a drone.
02:35
We will compare it in the background to
02:38
other types of imagery like
02:41
actual commercial planes or birds.
02:44
It will cancel them out and leave us only with the image.
02:48
Once the alarm is triggered and we have
02:51
an actual alert of a drone, we will deliver
02:53
an automated message via SMS, email,
02:58
or even an acoustic or light alert
03:00
within a center of security operations.
03:04
Any positive alert of a drone will immediately flash
03:08
the contour of the box in red, so we will not only
03:11
get the textual alert that there is an alarm,
03:14
but we will get the red flashing.
03:15
We can also put acoustics into that, and we will get
03:18
an alarm actually sounding.
03:20
And the red dots are definitely on the wireless
03:24
or the WiFi sensor when we identify a MAC address
03:27
or an SSID that's correspondent to a drone.
03:31
So it will make the difference between a mobile phone,
03:34
a smart phone, a wireless router, and an actual drone
03:38
from any company that develops drones that use WiFi.
03:42
We will frame them in green when they're not a drone,
03:46
like birds, like foliage moving in the background,
03:49
or a flag, and anything that's positively identified
03:52
as a drone will be framed in red.
03:55
We can identify the good drones by basically feeding them
03:59
into the library and labeling them on whichever aspect
04:03
we define as good drones or friendly drones.
04:06
Label them as a police drone or a fireman drone
04:09
or a media drone in different types of events
04:12
and cases where they will be flying.
04:15
The user interface will give us an inventory of
04:18
all the sensors that are active.
04:20
In this case, we will have two or three sensors
04:23
that are being visualized at the same time,
04:26
and we have different windows for each one of them.
04:29
So we will get the live video feed on one window,
04:32
and on the next, we will see the different acoustic
04:35
levels that are being sensed or recorded at the time.
04:38
Everything is automatic from the sensing to the alert
04:42
and even automated recordings of the events
04:45
for forensic applications, for forensic use,
04:49
is done automatically.
