#107 Robert Rose, Reliable Robotics: Congressional testimony and conveyor belts in the sky

In this episode of The Vertical Space, Robert Rose — co-founder and CEO of Reliable Robotics — returns fresh off his testimony before the House Transportation and Infrastructure Subcommittee on Aviation. Robert shares what most people misunderstand about FAA certification: the regulator isn't there to coach you through it, they are calling balls and strikes. He explains why Reliable has spent eight years building autonomous systems within existing regulations rather than waiting for new rules, how they convinced the FAA that zero-visibility automated landing standards can scale from wide-body jets down to Cessna Caravans, and why the cargo-first narrative that dominates autonomy discussions is largely a regulatory myth.

The conversation also covers Reliable's Pentagon contract to deploy autonomous cargo aircraft for contested logistics in the Indo-Pacific — what the military calls building conveyor belts in the sky — and why military logistics demands commercial-grade safety in ways most people don't appreciate.

Key Topics

• Why FAA certification works like an umpire calling balls and strikes — and what that means for autonomy companies
• The three phases of certification: legal basis, compliance plan, and execution — and where most companies get stuck
• What Robert told Congress about FAA morale, the reauthorization act, and what is actually improving
• How Reliable built its entire regulatory strategy on precedent rather than waiting for new rules
• How CAT IIIB autoland standards for wide-body jets were adapted for a Cessna Caravan
• The Pentagon contested logistics contract and why military cargo demands civil-level safety
• Why the cargo-first narrative in autonomy is mostly a regulatory myth
• How solid-state radar became an unexpected multi-billion dollar opportunity for existing airlines

+ Read Full Transcript

Robert Rose: The Air Force really likes using commercial technology because then they don't have to dig into things at any level of depth. They can rely on civilian certification. This was part of the promise of the 767 and the KC-46, for example.

Jim: Hey everyone, welcome back to The Vertical Space. Today we're joined by Robert Rose, co-founder and CEO of Reliable Robotics. Robert recently testified before the House Committee on Transportation and Infrastructure Subcommittee on Aviation at a hearing titled "America Builds: The State of the Advanced Air Mobility Industry." We thought he was terrific, so we had him join us to expand on that testimony, share his perspective on where the Advanced Air Mobility industry stands today, and give us an update on what's happening at Reliable Robotics.

Jim: Is there anything that very few in the industry agree with you on?

Robert Rose: There's a misconception I've encountered repeatedly about how certification works — this sense that for advanced automation, safety-enhancing technologies, and autonomy to be certified, we need to wait for the regulator to figure this out. That's just not the way the system works. The FAA is like an umpire at a baseball game. The umpire doesn't give you any coaching advice — all they do is call balls and strikes. If you develop a new technology, you can't just show it to the FAA and say "what do you need to see?" They're just going to say, is it safe or is it not safe? When the game changes — when you're talking about new technology that's never been put into the airspace before — it's still your job as the applicant to explain to them why they can make that call legally. This is something I've encountered repeatedly all across the industry.

What Actually Makes Certification Hard

Luka: Certification as a formal process is not necessarily the bottleneck that people think it is, but rather the substantial work required to make certification possible. What's the hardest work that has to happen before certification can even really begin?

Robert Rose: Certification is two or maybe three phases. The first is establishing the legal basis for the certification activity — what gives the regulator the authority to engage in this process in the first place? For technology that's been certified before, you can pull guidance off the shelf. But for new technology, you have to go through that process. Then you decide how you demonstrate compliance with the regulator. Then the third phase is you just actually do it — you run through and show compliance. Where people get tripped up is that in most other industries, you're only doing the third step. You're building a thing, testing it, calling it good — self-certifying. You're not first thinking critically about the legal basis for the equipment or thinking in advance about how you're going to demonstrate compliance.

Luka: So you also have to explain to the regulator how they're supposed to do their job — agree upfront on what the strike zone is.

Robert Rose: Exactly. That's not that controversial. I encounter it quite often and it's a message I want to get out there more.

Testifying Before Congress: FAA Reauthorization, Morale, and What's Improving

Jim: You recently spoke before Congress. Why were you there and what were the highlights?

Robert Rose: We went through this process over the last several years with the FAA 2024 Reauthorization Act — one of the most significant reauthorizations in a generation. Our team has been working with the FAA on this certification project going back to 2017. We've learned things about what works really well with the FAA and what could be improved. Some things that need improving require support and monetary support and direction from Congress. I was asked to testify to provide a status update on how that law is being implemented and whether we're seeing positive changes. The gist of my testimony: things are going pretty dang well. The legislation has moved the FAA very positively in a good direction. The FAA went through a bit of a dark time over the last several years — morale was quite low. But they're on the uptrend now. The attention from Congress through the FAA Reauthorization Act and the passing of legislation around ATC modernization and funding has done wonders to improve morale and increase their ability to make decisions and expedite decision-making. What we've noticed over the last two years has been an acceleration in challenging work — issue papers moving forward, standards moving forward. It's all been very good.

Luka: Was your motivation more aimed at unlocking funding, regulatory clarity, or just overall institutional confidence?

Robert Rose: You don't go into a meeting like that with an ask — that's in poor form. The ask is made outside the meeting. In that meeting I was trying to highlight what's working really well. Behind closed doors I'm a little more direct about what I think we should be focused on next. The size of the ask has really decreased over the years because things are moving in the right direction and it's really more around subtle adjustments around the edges. One thing I tried to highlight was that the FAA has been a great partner in moving forward the detect-and-avoid standards. RTCA through Special Committee 228 published standards for Detect and Avoid utilizing radar and ACAS X algorithms. The initial draft was really intended for operations in Class Alpha over 18,000 feet. For an aircraft like what we're working on — certifying first on the Cessna Caravan — those standards were kinda overkill. So we worked with RTCA to enhance that standard to accommodate smaller vehicles below 18,000 feet. The FAA has engaged with us on the adoption of the new iteration of DO-365 and DO-366, and around the same time as the testimony they released the changes to these TSOs to the public for comment.

Why Reliable Robotics Is a Different Kind of Company

Luka: When Congress is asking how AAM is doing, do they lump eVTOLs and autonomous aircraft and drones into a single AAM bucket?

Robert Rose: Reliable Robotics is a bit of a different business and technology project from everybody else. I am not aware of anyone else focused on how to certify safety-enhancing components that ultimately lead to a pilot not needing to be in the cockpit anymore. The frustration is there's not a collection of companies to lump us with other than other people doing new things in the sky. If I had to lump us, we're a lot closer to a traditional avionics company just continuing on this trajectory of improving safety. It's just we have a different ultimate end goal. Congress has a single unit — AAM — but there are a number of very enlightened members that understand the difference. When you walk in fresh, they see AAM and immediately think "you're that company working on the electric airplane." No sir, it's a Cessna Caravan — it burns jet fuel, but it doesn't have a pilot on board. They go, "what?" And then you spend four or five minutes explaining the difference and people go: "Oh, this makes a lot of sense. Why haven't I heard of this before?"

Jim: You made a comment where autonomy is ready now. Talk about what's ready now.

Robert Rose: With Wisk — mad respect for everyone else in the industry — but they're tackling a very different problem. IFR operation of a helicopter all the way to the surface in zero-visibility, total whiteout conditions is exceptionally rare outside of the military. There's not really precedence for that. So with Reliable, we're a fixed-wing airplane landing on a runway, and there is precedence for doing zero-visibility landing — CAT IIIB, CAT IIIC. There's already minimum operational performance standards published for how to do that. We come along with a Caravan — yeah, it's a Part 23 aircraft, but we can pull from the Part 25 standards. It's an inch above what you've already done previously. Autonomy is solved from the technology side — we know how to do this. The military knows how to do it. And from a regulatory standpoint, Reliable has solved it. We have worked with the FAA for the last eight years defining specifically how you take those large aircraft standards and pull them down to small vehicles and extend it into other domains. Automated takeoff is kind of just the reverse of the auto landing, pulling from those standards. Then we pushed it into automated taxi standards. There are a few things we made up — the FAA didn't have a standard for how closely you need to track the taxiway yellow line, but we came up with one and an integrity metric based on autoland and Part 25 precedent.

Reliable's Regulatory Strategy

Luka: How does progress happen more commonly in building autonomy — by adding capability or intelligence, or from constraining the problem to a point where you earn trust?

Robert Rose: What Reliable Robotics has chosen to pursue is what is possible within the existing regulations. The technology is not the hard part — we can science this all we want. The hard part is what's the legal authority for the regulatory body to certify it? And then how do you demonstrate that you comply? That vastly constrains the solution space. When you start from the regulations and flow things down, you start to exclude solutions that utilize non-deterministic algorithms or very large neural networks. There's currently no precedence for that — no standards. Academia doesn't even have consensus on how to do that. So you eliminate camera-based computer vision with neural networks, using neural networks to process lidar data, things of that nature. With detect and avoid, we eventually concluded that the only way to do this would be to have an active airborne surveillance component that emits radar in the nine-gigahertz spectrum. Once you have this requirement set, it's basically just physics from that point.

Luka: Did you at some point have that architecture?

Robert Rose: In the very early days. We had a document called MVA — Minimum Viable Autonomy — that outlined a multi-step progression to realize an autonomous aircraft. We kept adding to it. Every time there was a new problem, we'd add a row and describe possible technical solutions. As our conversations with the regulator matured and our understanding of the current regulations matured, we started going back through and crossing stuff off. This is obviously never going to work because a software algorithm — there's no way to prove through analysis that it meets a level of integrity. You can't prove that it's 99.999% accurate or whatever the requirement happens to be. For hardware components that rely on highly non-deterministic or environmental characteristics — very hard to certify.

Luka: Where else does constraining the problem to meet the regulatory framework show up in practice?

Robert Rose: One thing we held fixed in our requirements: we wanted approvals to fly anywhere, everywhere, all the time, in any condition. We didn't think the business was viable if we only had approval for one or a finite set of routes and every time we needed to expand that, we'd have to go through another approval process. In the US, the FAA has largely moved away from risk assessment-based approvals for larger vehicles. Reliable has been seeking approvals where there are no special conditions or limitations on operations, approved for use in all 48 states plus Alaska. It took years. Maybe we could have fought the FAA and lobbied Congress for a new part to make exceptions, but our belief in the early days was that wasn't likely to be a successful approach.

Update on the Autonomous Caravan Project and Contested Logistics

Luka: What is the status of the Caravan project? How has it evolved since we last spoke?

Robert Rose: We're at a point now where the approvals for this first Caravan STC are in place — the FAA project approvals are signed. Now we're executing — heads-down engineering focus mode, trying to deliver all the evidence to the FAA that we comply with these plans. The big thing that's changed: the Department of War has now become extremely interested in autonomy. They recently consolidated their priorities down to six items and listed contested logistics as one of them. Admiral Paparo, who runs INDOPACOM, testified before Congress about needing unmanned commercial airlift capability in the Indo-Pacific — now, he needed it yesterday. Shortly after these announcements, we won a contract with the Department of War to deploy our system to the Indo-Pacific region. The plan is to deliver this capability by the end of this year. It's the same technology going through the FAA certification process.

Conveyor Belt in the Sky

Jim: Tell the audience about contested logistics and what options the Department had beyond Reliable Robotics.

Robert Rose: The Indo-Pacific region — the military talks about the tyranny of distance and the need to control certain key points in the supply chain to resupply forward operating bases. They believe this logistics infrastructure will be contested. There's a concept called Agile Combat Employment — it's part operational strategy in how they run their resupply network and also the technology utilized in that network. We have a finite number of C-130s and C-17s to haul gear around the Pacific — they're very large and complex machines to operate. A solution like ours — a smaller unmanned aircraft — could be flown at much higher frequencies with greater risk because it's unmanned. For the price of a single C-130, you could have five or ten Caravans bouncing back and forth between airports continuously. The Air Force calls this a conveyor belt in the sky. You can dynamically reposition because you don't have crew limitations. I wish we'd gotten started on this a lot earlier, but here's where we are today.

Military Logistics Still Needs Civil-Level Safety

Robert Rose: There's a misconception that the DOW is a very different risk posture. True for portions of the DOW mission — weapons engagement systems. But logistics is different. Logistics starts in a civilian operation domain. When we take our airplane to Guam, it's not acceptable to crash on Guam. Very low risk tolerance there. So we need to follow many of the same standards. It would not be acceptable for our aircraft to crash on a runway even in the heat of war if the enemy is miles away. Taking out a runway or barracks next to a runway is an unacceptable level of risk. The Air Force really likes using commercial technology because then they don't have to dig into things at any level of depth — they can rely on civilian certification. That's part of the promise of the 767 and the KC-46.

Update on Reliable Robotics and the Industry

Robert Rose: I'd go back and emphasize the FAA Reauthorization Act and the restructuring occurring in the FAA. Having leadership in the FAA that's not in an acting capacity very much benefits the industry — there's a much greater chance of moving things through the processes quickly and empowering people to make decisions. The big thing that's changed for Reliable: we're on the home stretch now with our final sets of approvals. Department of War contracts are also new. Since we last spoke, this radar technology — we've seen tremendous inbound interest in utilizing it on existing Part 25 platforms. We're now working on a Part 25 version of this radar.

Robert Rose: Reliable is developing an electronic scan array antenna radar — solid-state technology. Radars in commercial aircraft today are gimbal-based with moving parts and are relatively slow and low accuracy. An ESA has no moving parts. We can electronically steer the beam and scan the sky in front of the aircraft up to 100 times faster than existing systems at much higher resolutions and accuracies. It has utility for collision avoidance, weather, high-altitude ice crystal detection, wind shear detection, turbulence detection, and we can put a low-power mode that enables collision mitigation on the surface — scanning the airport environment in front of the aircraft like the ultrasound sensors in your vehicle. The operators and OEMs we're talking to aren't ready for autonomy yet — it's scary. But the problem right in front of them is they want to replace their existing radars with more reliable technology. The resolvers and other moving parts in those systems create maintenance nightmares. If you have a weather radar that can scan the environment in front of you in three or four seconds instead of a minute or more, you can request a diversion or different heading from ATC much sooner — fuel savings and improved dispatch reliability.

Cargo-First Mythbusting

Luka: There's a hidden tension in the cargo-first approach — either cargo is a rehearsal where you have the same system and same safety case as passenger aviation, or cargo lives in a completely different regime with different risk tolerance and operational tempo. Do you see this tension?

Robert Rose: When we started the company, we thought flying cargo out in the middle of nowhere would make this project somehow simpler with the FAA. What we learned is there's no regulatory way to take credit for that. The process is Part 23 and Part 25 aircraft certification — it's all about the weight of the vehicle and maximum gross takeoff weight. There's no mechanism to really take credit for cargo versus passenger or rural versus urban. However, we stayed on the Caravan and cargo because of pull from the market — there are more operators ready for autonomous Caravans today than operators ready for autonomous passenger-carrying aircraft. Cargo is a stepping stone in the sense that the public will take notice and it will go a long way to improve public trust. But there's no mechanism to take credit for that with the regulator.

Luka: If you were starting Reliable today, would you make the same sequence of decisions?

Robert Rose: My co-founder and I keep coming back: yeah, this is the right plane. It makes sense. It fits within the right segment of Part 23. The use case is the right segment. We've got customers and operators lined up who want it. If this technology were certified tomorrow, we'd have hundreds of these things in operation like that. The holdup right now is us — we just need to actually deliver.

Scaling Autonomy: Certification Tooling, Digital Flight Rules, and Closing Thoughts

Luka: As you look into the future, how dependent are you on things like automated flight rules with respect to aerospace integration?

Robert Rose: Our strategy is to develop technology that integrates into the airspace system we have today. As we lower the cost of operations, we're going to increase the number of aircraft — increase frequency and then number of aircraft — and we'll start to hit constraints in the airspace in the form of congestion and controller workload. That's going to be a problem, not for 10-plus years, but it will become one. That's where our efforts with Congress and the FAA around digital flight come in. Surveillance — ADS-B is table stakes, we're big supporters of ADS-B in and out mandates. Communication — furthering digital ATC communications and CPDLC. We should push CPDLC out everywhere. When you get your instrument rating, people should just be doing this all digitally. Third: better real-time digital voice communications between pilots and controllers no matter where the pilot happens to be. We've been pushing for private operators to have access to the FAA's VoIP network so we can decongest the VHF AM radio system. If you can do those three things, it's not that much further a stretch to move into digital flight rules — what Wisk calls Automated Flight Rules. But you gotta have those three building blocks in place first.

Robert Rose: We've invested heavily since the early days in a systems engineering tooling process framework for managing all the regulatory material. We call it internally T-TREQS — traceable requirements. We've ingested 25,000-plus pages of regulatory guidance material into a structured model that maps onto our product breakdown structure, functional breakdown structure, system design, and test evidence. This is what's going to allow us to get onto the next vehicle rapidly. When the next vehicle comes along, we'll take that vehicle architecture, adjust our system architecture, and map our functional breakdown onto it so we have traceability back to the requirements. A study we did last year for the Air Force on the KC-135 showed this tool gives us a pretty significant competitive advantage in ability to scale.

Robert Rose: For Delta — two things. Automated landing: you want to increase dispatch reliability. You want weather to be less of an issue. When the weather at San Francisco gets below minimums, they switch to single-runway operations and you're burning a ton of fuel waiting for your slot. Automated landing can be a huge benefit for increasing operational efficiency and improving dispatch availability. The second is the radar — ground surface collision mitigation is top of mind for many operators. Shutting down a runway or taxiway means thousands of people inconvenienced and real costs in the millions and billions of dollars. Our radar can scan for obstacles including things that might not have transponders — the luggage that fell off the bag cart sitting on the taxiway for some reason.

Robert Rose: Whatever your interest level is, please start with safety. Our industry is amazing but safety is our lifeblood. New types of digital services can be created — for air navigation service providers, airlines, airports, vertiports. Building the digital backbone, integrating those technologies together, working with government and non-government agencies — that's really important. Working with ICAO to ensure the technologies come to fruition and work seamlessly together is imperative.