Successful Remote-Piloted Flight For Cessna Caravan


California-based Reliable Robotics announced today (Dec. 6) it has completed an FAA-sanctioned uncrewed flight of a Cessna 208B Caravan at Hollister Airport (KCVH) just south of San Jose, with a remote pilot in control from the company headquarters in Mountain View, California, 50 miles away. The Caravan took off from Runway 31 and entered a closed traffic pattern for landing. Designed for autonomous flight, the company’s technology enables remote operation through all phases of flight, including taxi, takeoff and landing.

Reliable Robotics said its technology is “aircraft agnostic” and has multiple layers of systems redundancy. The system incorporates “advanced navigation technology” and is also designed to prevent controlled flight into terrain (CFIT) and loss of control inflight (LOC-I).

One target market is enabling cargo flights to remote areas. The announcement noted, “The Caravan, and other regional cargo aircraft like it, serve an essential role connecting communities and businesses across the United States and around the globe.” Chris Hearne, Textron Senior VP of Engineering and Programs, said, “Reliable’s successful flight of an uncrewed Cessna 208 Caravan represents a milestone for the industry in bringing new technology to aviation.”

Last August, Reliable Robotics announced it had completed simulations and flight tests as part of the FAA’s Urban Air Mobility Airspace Management Demonstration. The simulations involved integrating remotely piloted aircraft into congested airspace, according to Reliable. The flight trials took place over the course of a week in northern California.

Today’s announcement also revealed that Reliable has been working under multiple contracts with the U.S. Air Force since 2021 “to leverage the significant progress on remote piloting for the Cessna Caravan to jointly examine how this commercially derived technology can be applied to large multi-engine aircraft for cargo logistics, aerial refueling and other missions.”

Mark Phelps
Mark Phelps is a senior editor at AVweb. He is an instrument rated private pilot and former owner of a Grumman American AA1B and a V-tail Bonanza.


  1. Well … I lied. I said the V-22 Osprey was the ‘only’ airplane I wouldn’t fly on. Now — with this news — there’s two … any airplane without a pilot.

    AND … what happens when the GPS constellation goes down or is compromised as has recently happened or the usual ‘three things went bad’ accident scenario hits … THEN what? It’s one thing to have a red button in the cockpit that a passenger can push if a pilot conks out; it’s entirely another when you never had a pilot at all. And aerial refueling without a pilot … is it OK if I remain massively skeptical?

    WHY is this nonsense necessary? Have our yout’s gotten SO lazy they don’t want to learn to fly? Geesh.

  2. This could be the start of a new legitimate career for Trevor Jacob once he is released from prison.
    Red Bulls need not apply.

  3. The problem isn’t “what to do when this or that happens” and a pilot is needed to intervene. If you can imagine a contingency you can plan for it. Most accidents are not pilots flying crippled airplanes ( certainly not turbine aircraft) but pilots flying perfectly good aircraft into the ground. Drones would not do that, and they wouldn’t do that at a rate and reliability that would and should shame the pilot community. Drones wouldn’t load their families on an overweight and underpowered Cherokee at high density altitude and crash it. Drones wouldn’t fly into weather it couldn’t handle because a drone doesn’t have to prove it “still has it” and can get in because it always had before.

    A drone will probably operate so conservatively in pursuit if a 100% safety record that it will be as unreliable a means of transportation as air travel should be.

    • +1
      This is happening, Old Pilots. Get used to it, or don’t. But please, stop trying to come up with reasons that you don’t think it’s safe, or that it won’t work. You obviously don’t know enough about designing modern electronic systems to even think of all the potential problems, let alone all the potential solutions to those problems. Just keep thinking that piloting is something that only a human – someone as smart, and experienced, and as level-headed as you – can do well.
      By the way, unmanned military drones have been sharing our airspace for years. You don’t hear about those just falling out of the sky, do you?

      • So,… you totally trust an autopilot? Hmm… yes I am “old” with 27,000 flight hours. I have seen many times where autopilots (even on newer planes like the B-787) have just decided to go awry and required human intervention on the spot. Seems like a drone pilot in a comfy room somewhere might just get to watch his plane while it crashes. The 787 has three autopilots and plenty of redundancy but it still failed requiring my human hands in the cockpit.

      • Your condescending all knowing tone is uncalled for. Engineer or pilot knows best- age old question…

        The military still has issues and no we don’t hear about it.
        The experienced pilots have seen the issues with so called fail safe systems. They aren’t often but do happen.
        This is great technology. I just don’t think it’s mature yet. I won’t get on a pilotless aircraft yet. How many have ridden in pilotless cars??? Yes it’s coming- just slowly …

      • I’m with Larry S, Raf, and the other pilots on this one: not with me as a pax.

        What makes you think we would have heard about any autonomous drone “deviations” from the military? They have huge swaths of airspace in which to test their toys and retrieve their failures without the public knowing.

        I live (and fly a chopper) five miles from one of the BVLOS test sites for the FlyTrex drone delivery service. Those sky-mines are completely invisible to aircraft, but deadly if you hit one. Supposedly they are “developing” their S&A technology, but meanwhile they are delivering life-saving hoagies to ten thousand oblivious, but technology-obsessed citizens.

        I was living only five miles from where that Tesla failed to distinguish between the color of the side of a left-turning semi-trailer and the sky behind it. The idiot Tesla passenger-in-the-driver’s-seat was watching a video at 65mph until he was decapitated. That Tesla was tested out the wazoo and approved by NTSA and all the other alphabets and sold on the open market. The car’s instruction manual did say that the driver should constantly pay attention to the road.

        I have no interest in being a guinea pig in an experiment in technological Darwinism.

      • And this, ladies and gentlemen, is the perfect example of why this is a bad idea.
        Although he’s right that it’s coming, the problem is right in the post:
        Programmers and engineers who think they can predict all possible failures and account for the in advance.
        Has never been done yet with 100% success, and it never will be.
        One may argue that it can match or exceed the success rate of human pilots, but it is anything BUT a sure thing.
        Hubris, whether that of the pilot, or of the designers, is the actual enemy.

  4. Another “ding” for aviation.
    Since the FAA is lowering it’s standards for air traffic controllers and rushing future pilots through training, somebody has to automate the planes.
    And these stupid electric flying taxis these companies are trying pushing.
    The Wright brothers would be turning over in their grave.

    • The Wright brothers would be leading the pack on all of this! They had vision, and they had the mindset to do things that no one thought could be done. Why would they have stopped imagining the future, like so many readers of AvWeb seem to have done?

  5. Air-Tesla.

    Change is hard to accept for all of us. These aircraft will first be augmented with human monitors. After 1000’s of hours of proving runs the monitors will no longer be needed.

    Fully autonomous passenger (non pilots ) carrying drones will beat these Caravans to market. The Chinese will capture the drone market first because the have a lust to win, and they have an empty sky. Their skies are not full of free market birds such as 172’s, Citations , and 737’s. Makes it easier to fill the sky with drones starting with a clean slate.

  6. Like it or not, its coming. And I say that as a cargo pilot. As long as the data shows an equivalent (or better) level of risk than traditionally crewed aircraft, the emotional argument evaporates. But the emotional argument WILL hold weight with the people carriers. After all, the paying passengers get a vote regardless of what the data says. However, gathering sufficient data from actual unmanned flights in order to make a meaningful comparison will take a long time. And then lawyers and insurance underwriters will have to be convinced. I’m pretty confident my seat will be safe for the next 10 years.

    • Why was Kroger necessary? The little mom-and-pop grocery stores were doing just fine, weren’t they? Why was Amazon necessary? Main Street USA and Malls of America were providing us with everything we could ever want in the retail world, weren’t they?
      It’s not about “necessary”. Progress rarely is.

  7. These comments are EXACTLY what I expected them to be. ‘Thats great, but what happens when X quits d’ja ever think of that???!!!!’
    Umm… no, no engineer ever thought of that. They’re smart and all, but nope, they never considered that critical components in their system might fail so they never bothered to design one or more fail safe mechanisms for it. Good thing you came along to show them how massively they just screwed up in their design because no engineer is aware that things break sometimes./s

    Love it.

      • So I guess you’re saying the engineers responsible for the three Tesla deaths (so far) in Florida alone, “thought about that” and still didn’t design enough “fail-safes” into the car? That sounds like culpable negligence to me. Hope the families sued their CAD-asses off.

        • The Tesla itself, and even more so the marketing and mindset of the owner, have been and continue to be a serious safety risk on our roads.

          To see this philosophy extended to the skies is tragic.

  8. I have a slightly different reason to ask “WHY”. I am having trouble understanding the economics.

    The pilot is the lowest cost per hour of almost any “component” in a fly-for-hire airplane, even a Caravan. The hourly depreciation cost of the engine, the airframe, maintenance cost per hour, fuel burn, are higher by a huge margin that the hourly cost of a pilot. The economics leave me scratching my head.

    It gets even more lop-sided when you look at an airliner. The hourly cost of flight crew at say $400,000 salary divided by 1200 hours is about $333/hour. While the fuel and engine hourly cost alone is about 20 times that amount.

    So why all the focus on eliminating the least expensive part of the flight?

    • Eliminating the pilot is currently needed because there isn’t a surplus of pilots to fly these smaller, not well paying, jobs.
      You could also look at it this way. If we DID have enough pilots, who would be flying these types of jobs? The least experienced/capable people. So the computer might be better at this, even with the possible failure points.

    • The only reason is that the pilot, in Mountain View, can jump on his e-scooter and be home for supper and bed, while the poor little aeroplane has to sit in the cold, dark and windy mountains for the night till the pilot gets in to the office, charges the scooter, drinks his coffee and finally decides to go to work….

  9. One of the big reasons developing autonomous flight technology is attractive to researchers is something pilots probably don’t want to hear: Flying a plane is simpler than driving a car. The physics of actually operating both are similar in magnitude and relatively easy to program for.
    Where The number of hazards faced by an autonomous car border on the infinite.

    • To be safe all this requires avionics that are more redundant than what we have now. Modern airplanes can and do experience new failure modes where a single faulty bit on some inter-box communications bus can have massive effects (like multiple indications going blank or bonkers). A remote pilot lacking sensory input may find it harder to sort out what’s wrong (and certainly wants a bulletproof comlink).
      The article doesn’t mention how „see and avoid“ is handled.

  10. Well, the un-editability of this site’s posting process strikes again, but I guess I got the main point across.

  11. Automation has been coming for many years, and the drive behind this is because airline management has always hated the pilots. The pilots are the only factor which they have no control over, once in the seat the pilots have total control which makes management seethe, not too mention pilot pay. Just imagine the millions (billions) management will stick in their pockets when they get rid of the pilots. As far as safety is concerned the only group really pushing safety over the years is ALPA. Pilots or automation there will be accidents, I have yet too see anything work flawlesly. Suck it up.

    • “Suck it up? On automated aircraft, I hope they come equipped with a ‘Suck it up’ button for when things get tough. Just imagine a cockpit announcement: ‘Ladies and gentlemen, turbulence ahead – please press the ‘Suck it up’ button for a smoother ride.’

  12. Imagine C-3PO flying a Sopwith Camel.

    I guess I’ve got a weird take on this. Seems to me where the robot should go in the equation is on the assembly line. Seriously, with the price of a caravan, is the wage for the pilot (most of which goes to the government rather than his plate in one way or another) REALLY the problem?

    Also, is the best way to upgrade a caravan actually to eliminate the pilot? It’s a pretty slow, inefficient, and old design.

    It seems obvious to me we have created policies that are leading to weird decisions in the not so free market like essentially robotizing antiques. We all then have predictable reactions, but we don’t really understand what’s really the problem.

    • I agree. I don’t want AI in an airplane design-compromised to require the stability a human-pilot demands. I want something more efficient that could only be controlled by a high powered computer.

      If you look at the Aviation Safety Network you’ll see half a dozen or more crashes just today of varying severity. Pilots do dumb things much more often than aircraft systems mess up. A drone won’t try to save a few bucks by running old auto gas in the tank just because he can. It wont rush through the test flight cards out of impatience and wreck the plane he just built. It doesn’t turn the fuel selector off after shutdown and disregard checklist useage and forget to turn it on with just enough fuel in the line to get to 200 feet.

      No, there might be arguments against drones ( and they really can’t share airspace with non-autonomous aircraft) but safety isn’t one of them.

      • Actually, you could create an app that would do all of the above.
        I think it would be interesting to build a learning model for risk/reward behavior, and then see what risky actions a generative AI would come up with. If there is no self preservation parameter, you could get something like
        “Attempt to land a fully loaded 747 in a Costco parking lot”. Probably not the sort of thing that would get a lot of speculative venture capital, but you never know.

      • Only thing is, all those crashing pilots are flying essentially antique aircraft. We really have no idea how safe airplanes would be if we hadn’t basically stopped improving them 50 years ago.

        We do know Diamonds are twice as safe as any equivalent pistons while being merely evolutionary improvements over the competition. We know that we have had a great leap in avionics in the last 20 years, and because of the cost of flying and FAA intransigence, we haven’t really taken full advantage of it in pilot training and practices.

        I’ll give the robot pilot idea credit for likely being able to improve safety faster than the current industry can improve human pilots. I blame the FAA primarily for holding back the human pilots though.

        How many pilots you know fly in an airplane as well equipped as that caravan must be? It’s essentially an experimental plane with avionics we cannot buy at any price.

        In fact, now that I think about it, the robot pilots aren’t likely going to have good RoI versus a single pilot because there will not be volume enough for the aircraft.

        • Other than the remote stuff in that FedEx Caravan, that’s how every FedEx Feeder Caravan is equipped. When I started flying them new in 1990 they had boots and a good panel and autopilot. Over the years we got GPS, radar altimeter, multifunction display and then full glass and TKS deice as the one in the video is pictured. I spent more than 25 years working in those. Most IFR aircraft working in part 135 are equipped like that.

  13. Not sure why this is a story, remote/tele-controlled aircraft have been in routine use since the Vietnam war 50-years ago. As DoD likely has millions of hours of operating experience on these platforms, it would be helpful to share the bad things that happen so the public can have a realistic set of expectations when the civilian versions start crashing and killing people. For whatever reason, autonomous vehicles (UAMs, aircraft, or cars) will be held to stricter safety records than those same vehicles operated by people. The GM Cruise RoboTaxi operation is a great example, yes fewer crashes than people driven cars, but driving over a pedestrian and dragging them is not acceptable for a machine. When the autonomous UAMs start killing passengers and people on the ground (‘urban’ part of UAM), the FAA will go into Boeing 737 MAX mode and ground the entire enterprise until 10 to n++ safety is achieved.

  14. Love it. They will be selling access to the right seat and PIC logging privileges in no time. Can you imagine the value of watching a computer complete a perfect flight ? If the guy in the box at Mountain View is a CFI he might give you the controls while checking you out for the BFR. Imagine being able to preview an approach with someone who’s done it a thousand times … Super exciting stuff and nothing but opportunities as I see it.

  15. The world today may have a total of One Million Pilots. The population of the world is over Seven Billion. That’s one pilot per 7,000 citizens. Autonomous Aviation will bring air travel to most of the 7 billion everyday on their schedule.

    Pilots don’t need to like the future of aviation but, the 7 billion people who would like to commute by air will. There is simply more money spent on Autonomous aircraft then on the psychological profile of the million pilots alive today. Don’t think it’s coincidence that pilots are asking the FAA to be the “Thought Police” and the new’s releases of all these companies autonomous projects.

  16. I think this is an awesome proof of concept and a step in the right direction.
    From an operations point of view, planes with such autonomous capabilities will most definitely still require a qualified pilot on board to monitor and take over if there is a situation occurs the automation is not properly (or at all) prepared to handle, so it’s really not that much different from airline flying today – except that the programming is handled remotely.
    So I think pilots will still be needed, just fewer of them.

  17. One would hope a monitoring pilot would do a better job of figuring out what the automation was screwing up than the Air France pilots did with their frozen pitot tubes.

  18. How comfortable are you with:
    A.) being driven while sitting in the backseat of an autonomous taxi
    B.) putting your children on an autonomous-driven schoolbus
    C.) being a passenger (out of the reach of controls) in an autonomous aircraft
    D.) having a “heavy” freighter aircraft fly over you & your family being autonomously flown
    E.) relying on RF-based communications with today’s “impenetrable” cybersecurity assuring control, guidance, and emergency communications

    I’m not there yet and likely won’t be in my lifetime. Perhaps following generations will build that trust, but as a child of the ’60s, I’m not there yet.