Silicon Valley Takes a (Careful) Step Toward Autonomous Flying

If Silicon Valley can fly as well as they can drive, we’re all in trouble.

The most dangerous items in vehicles are #1 the driver/pilot and #2 the drivers/pilots of other vehicles. Weather and mechanical failures are comparatively minor factors. It is easy to make an autonomous car with a very low risk of accident if there is no other car or any other traffic on the road. Now helicopters are difficult to fly due to the complex interconnection between rotation rate of the main fan, speed, wind, angle of attack, secondary rotor, and so on, but this is a space that machine learning techniques can understand fairly easily. In 3D space there are many more ways to avoid another flying item than on a two-lane road. The consequence of all this is that is likely significantly easier to make a safe autonomous helicopter than a safe autonomous car at present. At least a competent artificial co-pilot is probably likely in a not-too-distant future. This will be helpful but this will not help road congestion anywhere because flying helicopters or any other vehicle is going to remain many orders of magnitude more expensive than cars.

I'm an instructor at a local flight school with two Robinson R44s (the helicopter model featured). When people ask about safety I say "For $400,000 it comes with about 1/100th of the intelligence of a $400 DJI drone." Probably 90 percent of current light aircraft accidents could be prevented with the hardware and software that is common within drones. What would be a nice first step is a software assistant that can make a single-pilot operation as safe as an airline flight with two pilots working together with checklists. Probably 80 percent of current accidents wouldn't have happened if there had been a two-pilot crew in the front seats using the manufacturer's checklists. Why not a camera up at the dome light watching the same gauges that the pilot can see and saying "You've got the GPS programmed to fly to Atlanta, but you have only enough fuel to make it to South Carolina."

“There are many things that must come to fruition before autonomous aircraft start flying people” I think we may be a bit behind the curve on this one. https://www.bloomberg.com/news/articles/2017-09-26/dubai-stages-first-pu...

I went through flight training at a regional airline about five years ago. Some younger pilot mentioned being an airline pilot for forty years. I said, “There won’t be pilots in forty years.”. The room fell quiet. They knew I was right. Pilots will be replaced by the equivalent of a “physician assistant” qualified monitor that will be able to troubleshoot systems (programmers and the like). Pilots will be regulated to more of a private luxury for personal use only.

people may think how can a computer replace a highly trained, smart and educated pilot running a complex machine in a dangerous and difficult task? we don't even have autonomous cars yet. i think the task on the road is infinitely more complex with far more ambiguities and variables. in the air, pilots are highly trained and medically and psychologically cleared to fly a machine with prescribed regular maintenance. on the ground, any fool with $50 in their pockets can drive a 40yr old piece of junk around on our crowded streets full of potholes and pedestrians. the hazards in the air are typically weather or geologically related and predictable; while on the ground you also have to contend with a host of other human factors.

@Cade Metz, it would be good background material if The Times wrote an (or another article) on how Boeing, Airbus and others use autonomous technology in current day commercial jets. I have seen bits and pieces in other tech mags. Maybe these companies are emulating them or are all this "stuff" new? If new, how long will it take to market when you filter out all the optimism. The SkyRyse system may be the most viable because they use existing aircraft but it would still be a leap to use it over populated areas without some assurance it won't crash into someone's living room. Or crash into a vehicle if flown over streets. Not to veer off subject but how about if the tech industry end or manage the speed of this insane personal scooter use on city sidewalks? They are a real menace and noisy. I see joyriders, not commuters. I can visualize these personal helicopters and flying machines morph into live air battle video games...over urban environments.

I’m sorry folks but all this talk of autonomous flying machines and cars is at least 50-100 years away from happening if ever. There is just too much involved in manipulating a machine like an automobile or a helicopter for a computer to do the job. Yes I know about AI and all that but to teach a computer to make human judgements is impossible. What really bothers me about this kind of technology is that we are spending billions of dollars on it while people still starve. I just can’t fathom that anymore. If you don’t believe me then just ask Waymo who is the leader in autonomous driving cars and they are even admitting it may be ultimately impossible. To say nothing of the ‘leap of faith’ that is involved in letting a computer decide life and death decisions for you. You’ll see in the long run that this idea will go the way of so many other bad tech ideas over the years that never transpired.