The Next Competition in Autonomy May Not Be About Building Better Machines, but Better Environments for Machines to Operate Within
By Joe Cozart
The unmanned-systems industry still tends to describe itself through machines. A new aircraft enters service. A battery lasts longer. A radar sees farther. A drone receives permission to fly beyond visual line of sight. Another autonomous platform acquires another sensor, another processor, another mission. All of that matters, but it may increasingly describe only the visible edge of something much larger.
The more consequential development is that autonomy itself appears to be migrating away from the individual machine and into the environment surrounding it. That distinction changes almost everything.
For much of the modern unmanned era, the central engineering problem has been how much intelligence could be placed aboard a vehicle. Navigation, sensing, communications, power management, collision avoidance, mission planning and decision-making all had to travel with the machine because the machine could not assume that the environment around it would provide very much assistance. The autonomous vehicle therefore had to become increasingly self-contained.
That architecture made perfect sense while the surrounding infrastructure remained relatively unintelligent. But the environment is beginning to change.
Ground-based radar can now contribute persistent situational awareness over operating areas. Remote identification can become part of a common operating picture rather than merely a compliance requirement. Unmanned traffic-management systems can reconcile aircraft movements across shared airspace. Cloud platforms can distribute operational information among vehicles and operators. Batteries can report their own state of charge, temperature, health and faults directly into flight-management systems. Sensors increasingly communicate not merely with pilots but with software architectures capable of interpreting the environment continuously.
What appears, when each announcement is read separately, to be another incremental improvement in unmanned technology begins to look rather different when the pieces are assembled.
The battery knows itself. The aircraft knows itself. The airspace knows the aircraft. Radar knows what else occupies that airspace. Remote identification distinguishes cooperative participants. Traffic-management systems understand where those participants are supposed to be. Communications networks distribute the information. Artificial intelligence increasingly interprets the resulting picture.
The machine is still autonomous, but the environment is becoming intelligent too.
That may be the transition worth watching.
The analogy is not really aviation. It may be telecommunications. Nobody expects a smartphone to build its own communications network every time someone makes a call. The intelligence of the device matters enormously, but the usefulness of the device depends upon an enormous external architecture of towers, fiber, satellites, spectrum, switches, software, standards, power and regulation that already exists before the phone is turned on. The device enters the environment. The environment makes scale possible.
Autonomous systems may be approaching a similar threshold.
If so, the defining question of the next phase of autonomy will gradually shift from, “What can this machine do?” to something more consequential: “What can this machine do because of the environment in which it operates?”
That is a very different question.
It also changes the economic unit.
For years the object of fascination has been the autonomous platform itself. The aircraft received the attention because the aircraft moved. The drone received the investment because the drone could be photographed. The vehicle could be demonstrated, sold and deployed. Infrastructure is less theatrical, but infrastructure eventually determines whether thousands of vehicles can do what one impressive vehicle demonstrated.
That difference separates experimentation from systems, and systems are where industries become economies.
If autonomous systems increasingly depend upon persistent communications, surveillance, identification, navigation, compute, weather intelligence, cybersecurity, traffic management, command-and-control, regulatory permission and distributed sensing, then the strategic asset of the future may not be the autonomous vehicle at all. It may be the autonomous operating environment.
That possibility deserves particular attention in North Dakota.
For two decades, northeastern North Dakota has steadily accumulated many of the ingredients required for unmanned aviation: military missions, research institutions, test infrastructure, specialized airspace, private industry, regulatory experience, command-and-control capability, aviation expertise and Grand Sky. The understandable tendency has been to describe this achievement as the development of a UAS ecosystem.
That description may now be too small.
The next strategic opportunity is not simply to continue attracting unmanned aircraft companies or to become better at flying drones beyond visual line of sight. It is to recognize that the infrastructure assembled for aviation may represent the early architecture of something far broader.
North Dakota could become an autonomous operating environment.
That would mean thinking beyond individual platforms and even beyond individual domains. Aircraft would be one class of participant. Ground vehicles another. Agricultural machines another. Logistics systems another. Infrastructure-inspection platforms another. Emergency-response systems another. Defense systems another.
The state would not need to manufacture all of them.
It would need to create an environment in which they could operate.
That distinction is enormous.
The conventional economic-development model attempts to attract the company that makes the machine. The upstream model asks whether it is possible to build the place every machine eventually needs.
That place would require communications architecture capable of maintaining persistent machine connectivity. It would require surveillance and sensing that extend situational awareness beyond individual vehicles. It would require high-integrity navigation. It would require edge and cloud computing. It would require common operating pictures capable of reconciling participants across domains. It would require cybersecurity designed for machines that continuously exchange operational information. It would require regulatory architecture capable of permitting routine autonomous operation rather than exceptional demonstrations.
It would also require something considerably more difficult than technology.
Institutional coherence.
Autonomy does not naturally respect the organizational boundaries humans have created for it. The aircraft does not particularly care whether the communications network belongs to one agency, the radar to another, the road to a county, the airspace to the federal government, the electrical infrastructure to a utility and the computing architecture to a private company.
The machine experiences all of it as environment.
Humans experience it as jurisdiction.
That may become one of the great limiting factors in the deployment of autonomy. The technology may eventually be easier than the institutional architecture required to allow the technology to operate continuously across fragmented human systems.
That is precisely why a relatively small state could possess an unusual advantage.
North Dakota does not have to solve Los Angeles. It does not have to solve New York. It does not have to begin with ten million vehicles moving through an environment constructed long before autonomous systems were imaginable.
It can begin with space. Agriculture. Energy. Defense. Long distances. Manageable population density. Existing aviation infrastructure. A mature unmanned ecosystem. And institutions that are already accustomed to experimentation.
Those characteristics have traditionally been treated as geographic facts.
They may increasingly become technological advantages.
The statewide autonomous-vehicle concept therefore becomes much more interesting when viewed through this lens. The objective would not simply be to deploy autonomous vehicles across North Dakota. That would still place the machine at the center of the strategy.
The more ambitious proposition would be to build a statewide operating environment capable of supporting many classes of autonomous systems.
The distinction sounds subtle until its implications are considered.
A drone corridor serves drones.
An autonomous operating environment serves autonomy.
The first is infrastructure for a use case.
The second is infrastructure for an economy.
The first asks which vehicle comes next.
The second makes that question almost irrelevant.
If the environment exists, vehicles can come and go. Companies can come and go. Sensors can improve. Battery chemistry can change. Aircraft designs can evolve. Artificial-intelligence models can be replaced.
The platform survives.
This is the lesson embedded throughout the history of infrastructure. Railroads were more important than any particular locomotive. Electrical grids were more important than any particular appliance. The internet became more important than any particular computer connected to it. Cellular networks became more important than any particular telephone.
Infrastructure does something products cannot.
It allows products that have not yet been invented to become useful when they arrive.
That is why the most valuable autonomy infrastructure may eventually be infrastructure built before anyone knows precisely which autonomous systems will dominate.
The mistake would be to wait until the winning machines are obvious.
By then the operating environments supporting them may already belong somewhere else.
North Dakota has spent approximately two decades becoming unusually good at unmanned aviation. The upstream question is whether unmanned aviation was ever the destination.
Perhaps it was the first layer.
Perhaps the real opportunity is now becoming visible only because enough of the pieces have accumulated to see it.
The next competition in autonomy may not be won by whoever builds the smartest machine.
It may be won by whoever builds the smartest place for machines to operate.
And if that is where autonomy is going, North Dakota should stop thinking of itself merely as a place where autonomous systems are tested.
It should begin thinking about becoming the platform on which they live.
——— GMJoe™ ———
Clarity. Strategy. Sovereignty.™
Live Upstream.™
Books by Joe Cozart are available at: amazon.com/author/joecozart
GMJoe.org