Automotive Engineering - April 2024 - 28

Q&
A
Owl AI making better ADAS systems
with help from drones
Owl AI has steadily been improving the SWaP-C (Size, Weight,
Power and Cost) of the electronics in its thermal camera sensors
while also improving capability. The company's latest units have
three-and-a-half times more pixel density in roughly the same
size silicon area, as well as additional logic in the chip that now
uses a single board versus the three-board stack used previously.
We spoke with Owl AI CEO and co-founder Chuck Gershman
about this improvement and how it's opening up new doors for
his company. This is an edited transcript of our discussion.
Tell me about the improvements in
your sensor since we last heard from
you at AutoSens in the fall of 2023.
The whole SWaP-C concept and the
whole resolution concept were designed
for ADAS applications, specifically for
getting us from an L2 into an L3 and
anything in between. When we showed
this hardware in public at CES, we got a
secondary customer base that seemed
to be all over us. They weren't asking us
to build them something different. They
were just asking us, " Could we apply this
to our application, " and those were mini
drones. The combination of SWaP-C and
the resolution really got the attention of
the dual-use community. "
Are you hinting at the military realm
here?
It could be military. It could be surveillance.
It could be border protection. It's
anything associated with surveillance applications
or any kind of reconnaissance,
as well as search and rescue. We always
knew that this technology would be applicable
there, but we never actively marketed
it. Once we showed the form factor,
there was just this avalanche of interest. It
doesn't change our basic hardware design
premises in any way, shape, or form. That automotive size, weight
and power solution just happens to fit in small form factor drones.
Chuck Gershman
" How do you
generate production
revenue during
the qualification
timeline? "
Those benefits apply across the board. How did you react?
As a smaller company, we want to stay focused, but we were
kind of inundated with requests. So we are paying attention
to that.
How does that impact the automotive work?
One of the biggest negatives of trying to intersect automotive
product lines [as a tier-two supplier] is simply the time28
April 2024
line from product commitment to product deployment. It's a
long qualification timeline. How do you generate production
revenue during the qualification timeline? We had a thesis
that there were other mobility applications that could fill
that revenue timeline. One of the reasons we didn't overly
market into the drone space was that we might have to do
additional engineering. What we're hearing from the drone
guys is, " Just give us this, this and this as you have it, and
we'll take care of the additional engineering. Just give us
access to deploy cameras on this, " and
now you can fill your revenue gap. You
can create incremental revenue while
you support your longer-term qualification
plans for automotive.
If you just go back and look at what
happened in the lidar space, you'll see
this very problem. The only real lidar
companies that have survived are a
handful of Chinese companies where
the Chinese market bought up their
stuff and those who SPACed. The reason
that the ones that SPACed survived was
they had ample monies to cover this
valley of revenue depth between proof
of concept and production.
Finally, there are new regulations on
the way regarding nighttime pedestrian
protection, something a thermal
camera is obviously good at. The announcement
describing these rules
has repeatedly been delayed. What
are your thoughts on the new rules,
whatever they might be?
The more disruption there is in the
market and the more market uncertainty
that's created by that disruption,
the better it is for small companies. The
more that's understood and known as a
given, the better it is for incumbents.
So, quite frankly, NHTSA's playing around and screwing with
people's minds and being a little late is good for us. It creates
more disruption, more uncertainty. We like uncertainty,
as it relates to the market that relates to us. So, quite frankly,
I'm not sweating at all. And whatever they come up with,
whether it's more rigid or less rigid than what they have
there, it's still table stakes for us. It's easy to do. We've already
tested it against their original proposal, and we
passed all their tests with flying colors. So we're not really
sweating what the spec says.
Sebastian Blanco
AUTOMOTIVE ENGINEERING
Automotive Engineering - April 2024

Table of Contents for the Digital Edition of Automotive Engineering - April 2024
