WEBVTT

00:00:00.000 --> 00:00:03.000
Welcome to the Deep Dive. So today we're getting

00:00:03.000 --> 00:00:07.120
into something that is a pretty fundamental shift

00:00:07.120 --> 00:00:09.779
in logistics. It really is. For almost a century,

00:00:09.980 --> 00:00:12.400
right, moving packages was all about discrete

00:00:12.400 --> 00:00:15.660
math. You know, nodes and edges on a graph. The

00:00:15.660 --> 00:00:17.719
class that transportation brought. Exactly. But

00:00:17.719 --> 00:00:20.519
that model just completely collapses when you're

00:00:20.519 --> 00:00:24.329
talking about, say, Puck's drone fleet in Milwaukee's

00:00:24.329 --> 00:00:26.550
historic third war. Oh, it's a combinatorial

00:00:26.550 --> 00:00:29.010
explosion. You can't. The old models for trucks

00:00:29.010 --> 00:00:32.429
and set routes, they just, they break down. You

00:00:32.429 --> 00:00:34.630
have a massive swarm in continuous airspace.

00:00:34.770 --> 00:00:37.030
So it stops being a math problem. It stops behaving

00:00:37.030 --> 00:00:39.850
like a series of individual calculations and

00:00:39.850 --> 00:00:42.469
starts behaving like a physical substance, like

00:00:42.469 --> 00:00:44.710
a fluid. So they threw out the old formulas,

00:00:44.810 --> 00:00:47.729
things like the Bobula -Forsyth calculation.

00:00:47.929 --> 00:00:50.149
Yeah, that's just for optimizing a single object,

00:00:50.250 --> 00:00:52.560
like a bullet. Here they turn to the Navier -Stokes

00:00:52.560 --> 00:00:55.039
equations. To create what they're calling laminar

00:00:55.039 --> 00:00:57.619
logistics. Right. And what's so fascinating is

00:00:57.619 --> 00:00:59.820
they stopped trying to optimize the drone itself

00:00:59.820 --> 00:01:02.140
and started optimizing the flow of the whole

00:01:02.140 --> 00:01:05.819
fleet. The city is now mapped as a, well, a porous

00:01:05.819 --> 00:01:08.000
medium. Okay, let's unpack that central analogy

00:01:08.000 --> 00:01:10.659
a little bit. When you say pressure, you're not

00:01:10.659 --> 00:01:13.140
talking about air pressure. No, not at all. Think

00:01:13.140 --> 00:01:17.959
of it as logistical potential or demand urgency.

00:01:18.620 --> 00:01:20.920
So the Milwaukee public market, which is always

00:01:20.920 --> 00:01:22.540
busy, that's the high pressure source. It's the

00:01:22.540 --> 00:01:25.319
perpetual source, yes. And a customer order,

00:01:25.560 --> 00:01:28.379
say, from the Corcoran lofts, that instantly

00:01:28.379 --> 00:01:31.200
becomes a low pressure sink. The drone fleet,

00:01:31.480 --> 00:01:33.819
the fluid, just follows the steepest pressure

00:01:33.819 --> 00:01:36.439
drop. It's physics. And every fluid has friction.

00:01:36.680 --> 00:01:39.439
Exactly. Which here we call viscosity. Viscosity

00:01:39.439 --> 00:01:41.579
maps directly to the energy cost of complexity.

00:01:41.920 --> 00:01:44.079
So not just physical friction. No, it's operational

00:01:44.079 --> 00:01:47.180
friction. You lose energy from, say, Sharp, inefficient

00:01:47.180 --> 00:01:49.959
turns, those are high -viscosity movements, but

00:01:49.959 --> 00:01:52.359
also from regulatory constraints. Wait, regulatory

00:01:52.359 --> 00:01:54.620
friction? How does an FAA rule act like molasses?

00:01:54.900 --> 00:01:57.859
Okay, so take the FAA's Part 107 rule visual

00:01:57.859 --> 00:01:59.819
line of sight. An operator has to be able to

00:01:59.819 --> 00:02:02.120
see the drone. Right. We can assign areas where

00:02:02.120 --> 00:02:03.739
that's tricky, like near the Italian Community

00:02:03.739 --> 00:02:06.120
Center during a big event, an extremely high

00:02:06.120 --> 00:02:09.419
viscosity value. The solver's job is to minimize

00:02:09.419 --> 00:02:12.599
that viscous dissipation. So the path it calculates

00:02:12.599 --> 00:02:15.060
is naturally smooth, efficient, and compliant.

00:02:15.719 --> 00:02:19.159
All by itself. All by design. No need to pre

00:02:19.159 --> 00:02:21.919
-program no -fly zones in the traditional sense.

00:02:22.080 --> 00:02:24.240
I love that. And it's not just drones in the

00:02:24.240 --> 00:02:26.219
air. This is a hybrid fleet, right? Multi -phase

00:02:26.219 --> 00:02:28.699
flow. That's the term. Drones are the low -friction

00:02:28.699 --> 00:02:31.740
air fluid, the free stream. The ground droids

00:02:31.740 --> 00:02:34.319
are the ground fluid stuck in the high -friction

00:02:34.319 --> 00:02:36.419
boundary layer of the street. And they can switch

00:02:36.419 --> 00:02:38.740
between them. Yes, through the condensation maneuver.

00:02:39.180 --> 00:02:42.759
If the air viscosity gets too high, maybe a sudden

00:02:42.759 --> 00:02:45.280
gust of wind off Lake Michigan or restricted

00:02:45.280 --> 00:02:48.560
airspace, a drone can land on a droid at, say,

00:02:48.620 --> 00:02:51.680
Catalano Square. It transfers the package. So

00:02:51.680 --> 00:02:54.259
the mass moves from the air phase to the ground

00:02:54.259 --> 00:02:57.000
phase. And completes the last mile, even when

00:02:57.000 --> 00:02:59.960
the air is basically prohibitive. It's incredibly

00:02:59.960 --> 00:03:02.159
resilient. Which brings up the big question,

00:03:02.240 --> 00:03:04.479
right? If it's all based on physics, how does

00:03:04.479 --> 00:03:08.080
it handle, well... chaos what happens when a

00:03:08.080 --> 00:03:11.020
fire truck just you know blocks broadway yeah

00:03:11.020 --> 00:03:12.939
that's the beautiful part a traditional system

00:03:12.939 --> 00:03:15.020
would frantically try to recalculate thousands

00:03:15.020 --> 00:03:17.800
of individual paths it would be a mess a huge

00:03:17.800 --> 00:03:21.819
processing lag for sure but the fluid model it

00:03:21.819 --> 00:03:24.659
handles turbulence instinctively the blockage

00:03:24.659 --> 00:03:27.780
is like a wall appearing in the flow Physics

00:03:27.780 --> 00:03:30.680
dictates that a pressure wave instantly propagates

00:03:30.680 --> 00:03:32.960
backward through the fluid. So the drones react

00:03:32.960 --> 00:03:35.180
before they even get there. They slow down and

00:03:35.180 --> 00:03:37.520
divert organically based purely on the change

00:03:37.520 --> 00:03:39.819
in flow dynamics. It's true self -organization.

00:03:40.099 --> 00:03:42.800
And what about that FAA line of sight rule we

00:03:42.800 --> 00:03:44.719
talked about? How does it make a drone physically

00:03:44.719 --> 00:03:47.580
stop? That's handled by what's called a visibility

00:03:47.580 --> 00:03:50.840
scalar field. As a drone nears an area where

00:03:50.840 --> 00:03:53.599
that visibility factor drops to zero, the effective

00:03:53.599 --> 00:03:55.919
viscosity in the calculation approaches infinity.

00:03:56.439 --> 00:03:58.840
The fluid just freezes. It literally freezes.

00:03:58.840 --> 00:04:01.340
Flow into that non -compliant area stops. It's

00:04:01.340 --> 00:04:03.780
a physics -based enforcement mechanism. So the

00:04:03.780 --> 00:04:06.280
real genius here, it seems, is the shift in perspective.

00:04:06.500 --> 00:04:08.580
You're not tracking individual particles anymore,

00:04:08.719 --> 00:04:11.099
that Lagrangian view. Exactly. You stop caring

00:04:11.099 --> 00:04:13.280
about the particle and start managing the total

00:04:13.280 --> 00:04:16.139
flow field, the Eulerian view. Yeah. The system

00:04:16.139 --> 00:04:18.579
becomes completely self -aware. It's incredible.

00:04:19.220 --> 00:04:21.040
And if you think about the bigger picture here,

00:04:21.660 --> 00:04:24.360
the Navier -Stokes model takes things that used

00:04:24.360 --> 00:04:27.560
to be seen as just obstacles. like wind, or the

00:04:27.560 --> 00:04:29.939
urban canyon effect, where tall buildings degrade

00:04:29.939 --> 00:04:32.860
GPS signals. It maps them. Signal degradation

00:04:32.860 --> 00:04:36.779
becomes signal viscosity. It uses these forces

00:04:36.779 --> 00:04:40.199
not as problems to avoid, but as known terms

00:04:40.199 --> 00:04:42.819
in the equation to find the truly optimal path.

00:04:42.980 --> 00:04:45.019
Which leaves you with a final thought. It does.

00:04:45.459 --> 00:04:47.620
What does it actually mean for our infrastructure

00:04:47.620 --> 00:04:50.060
when physics problems from centuries ago become

00:04:50.060 --> 00:04:52.480
the most elegant, self -organizing solutions

00:04:52.480 --> 00:04:54.600
for our most modern logistics challenges?
