Why stop at self-DRIVING cars when we could have self FLYING ones?

A few weeks ago I had a very interesting Twitter conversation with Don Osborn and others, around this very topic. This is a summary of that conversation, followed by the reply I had no time to elaborate and share in that moment.

Self-flying cars might be safer than self-driving.

When Don Osborn made this hypothesis on Twitter, my instinctive reply was “self flying cars seem to have ALL the cons of self driving ones, PLUS the possibility to crash on me ALSO from above”. So I asked to explain that statement, and he kindly did it, starting with this tweet. Here is a synthesis of Osborn’s answer:

  1. My speculation on self-driving vs self-flying begins with assumption that the tech behind autonomous vehicles deals better with complex and ~predictable environment than with complex environment + random factors.
  2. [I also assume] that flying cars, as they get more numerous, will require a sensible and strict regulation as to where they fly (think air traffic control with road map in low altitudes). So, even tho there is an added dimension, this is a ~predictable environment (on a clear day).
  3. Compare with what self-driving cars must deal with on roads. A lot of visual input to parse, and the potential for misinterpretation (and fatal accidents, as we’ve seen). Add unexpected situations and random factors. Even with only 2 dimensions, arguably a less predictable environment.
  4. Three main issues I see (as a non-expert) with self-flying cars. First, as with any aircraft, highest risk times are take-off and landing. For self-flying car, highest complexity, lower predictability. Second is handling mechanical or digital malfunctions and safe emergency landings.
  5. Third is weather. Aeronautically, how much inclement weather or high winds could a flying car be designed to handle? Systemically, how would a stream of self-flying cars be able to handle, say, buffeting winds while maintaining safe distances between each other?

Self-flying cars: probably safer, yes, but are they needed?

First of all, thanks again to Don Osborn for explaining what he meant.

In a nutshell, I now believe that he is correct, but that there are much simpler and cheaper ways than self-flying cars to achieve the same results. To begin with, let me confirm point 2 above: in the real world, individual self-flying cars would have to be so regulated to not improve traffic congestion at all. We have known this for at least 20 years, from “Star Wars: Episode 1”:

Self-Flying cars instead of self-driving cars. Hmmm /img/coruscant_apartment_view.jpg
SAME traffic jams of the 20th century, SAME waste of time, just with a better VIEW

The only solution to this problem is, of course, public transportation, or at least systematic car-sharing. Individual cars are not sustainable, no matter what their fuel is, or how far from the ground they move. Safety-wise, aviation is very safe exactly because it still is public, heavily regulated transportation, by very few vehicles that are immensely easier to build and monitor than millions of hackable cars.

Speaking of safety, let’s not forget that once there are thousands of cars in the air, all it takes to cause havoc is getting just one cheap drone, carrying just a few grams of explosives, close enough to one flying car in those queues in the picture. So, self-driving “mass transit”, that is completely new generations of planes, helicopters, airships.. for short- or medium-haul commercial flight, yes. Individual flying vehicles, almost surely no, if you ask me.

In addition to this, self-driving cars seem really an overkill to me because:

  • cost: buffeting winds, or safe emergency landings, are VERY expensive to get right. To be safe enough, flying has to be much more expensive than moving on the ground
  • predictability does not need flying: in many places, flying cars would have to fly alongside unpredictable (flocks of) birds. But even without birds, if the problem is unpredictable events, it is much simpler and cheaper to physically separate cars from pedestrians on the ground, as I argued here than make them able to fly

The self-flying car that we could REALLY use

All in all, I would say that “self-flying” cars, would still be very useful and doable… but only if we compromise on very different “cars” than those imagined in the initial discussion. The self-“flying” cars that are much more feasible and may do a lot of good may be those that we already saw in… Star Wars, again:

Self-Flying cars instead of self-driving cars. Hmmm /img/X34-landspeeder.jpg

that is, nothing but plain, old “hovercrafts 2.0”:

Self-Flying cars instead of self-driving cars. Hmmm /img/hovercraft-2.0.jpg
the self-flying cars that we already (almost) have

What about reinventing hovercrafts, to make them not noisier than a traditional car, but able to drive themselves, on regulated paths separated from pedestrian ones?

If something like that existed, it would surely be way safer, and cheaper too, than any really flying car, no matter who drives it. And it would also need much less road maintenance (and bridges). What do you say? Should we start a crowdfunding campaign for hovercraft 2.0?