The alternative is autonomous AI trafficked flying cars that is networked with all commercial drone traffic, but that is 5-10 years away from being reliable and possible.
50-100 years is more likely. The complexity of automated low altitude flight is exponentially more complex than driving on the ground.
And they communicate in real time while individually checking their collision sensors for non swarm obstacles. The technology is 100% there, the flight capabilities do not matter nor the fact it’s a passenger vehicle; we have the algorithm and sensor packages right now to do what you think is a hundred years away.
You completely glossed over the small, agile part.
A drone can do a fucking barrel roll with zero impact on its flight capabilities or passengers. Have fun with passengers when an automated passenger drone needs to quickly change thrust direction because of wind shear.
That’s a matter of changing the PID compensations, it’s not a difficult problem. Do you think all drones can do a barrel roll? There are nonagile drones in operation right now, in swarm configurations. Not as large as a car, but again, that’s a matter of tweaking the PID, nothing more or less.
To put it another way you don’t restart learning how to drive a car each time you drive a new car. You tweak your internal pids to that of the new car. The difference is humans are stupid, slow, and have terrible coordination that can’t be immediately transferred over a fleet.
There are a whole lot less obstacles and unknowns in the air, as well as more planes (ha!) of separation available than a car.
When flying, you don’t really need to worry much about random pedestrians, for example.
If the entire system were completely automated, from the car all the way to ATC, and it’s essentially a taxi that you just tell what location to go to and it handles the rest… well that’s basically air traffic today minus the automated ATC part. (That isn’t to diminish pilots at all; just that I think it’d be a lot easier, in general, to replace a pilot than a taxi driver with automation. They’re both still extremely challenging problems.)
If you ignore take off and landing, birds, weather conditions, and everything else that makes flying more complex and dangerous than driving on the ground, sure.
And if you ignore construction sites with high cranes and not documented buildings. Or overland high voltag power cables, wind mills, hobby drones, and local variations of birds.
It‘s just taking the complex challenges of autonomous drivinf into the third dimension. Making it even more complex.
Other than the takeoff and landing, cars have to deal with those obstacles as well.
A computer running a citywide automated traffic system for cars would have all the same complexity, without the ability to separate traffic in three dimensions.
So yeah, if you ignore the parts that make it more complicated it seems easier!
If it is windy, it is far more complicated than driving on the road, especially in cities with taller buildings. Like not crashing into buildings is far harder than applying the brakes when there is ever changing wind shear that you can’t see. This applies to most days in most cities.
It isn’t that complex. The problem with current autonomous driving is the car can only infer what other cars are doing and what is around it, especially if we are talking about an autonomous car with idiotic vision mapping without lidar.
With a flying car that is directed by an AI that knows where every other flying object is, what every flying object is going to do, and the locations of every stationary object based on maps and lidar on the vehicle, you can keep collisions far less likely. Taking the human control out of the picture improves the conditions substantially.
I wouldn’t trust a flying car at all, but I would trust an autonomous one with AI ATC far more than an autonomous car going through a construction zone on a highway in a major city during rush hour.
The problem with current autonomous driving is the car can only infer what other cars are doing and what is around it
With a flying car that is directed by an AI that knows where every other flying object is, what every flying object is going to do
Spoken like someone from a culture where drivers are the only thing around because they have gotten so used to ignoring pedestrians, bicyclists, animals (wild or otherwise),… that might be found on the road and hasn’t considered what else might be in the air at all.
You would have a point if the flying car didn’t have to take into account the wind, updrafts, downdrafts, wind sheer off buildings, and a ton of other flying related stuff that helicopter pilots need to take into account which are barely noticeable to cars the majority of the time.
Then there is landing, the most dangerous part of flying. Imagine if the emergency braking in a csr needed to stop the car without spilling a liquid from an open cup.
Being on the ground is far less complex than flying, otherwise getting a pilot’s license would be easier than a driver’s permit.
With a flying car that is directed by an AI that knows where every other flying object is, what every flying object is going to do, and the locations of every stationary object based on maps and lidar on the vehicle, you can keep collisions far less likely.
This magically perfect AI would work on the ground too. I mean, it knows what birds are going do, why not people?
50-100 years is more likely. The complexity of automated low altitude flight is exponentially more complex than driving on the ground.
Drone swarms already exist. And they work better than self driving cars
Drone swarms are groups of small, agile devices with no passengers.
And they communicate in real time while individually checking their collision sensors for non swarm obstacles. The technology is 100% there, the flight capabilities do not matter nor the fact it’s a passenger vehicle; we have the algorithm and sensor packages right now to do what you think is a hundred years away.
You completely glossed over the small, agile part.
A drone can do a fucking barrel roll with zero impact on its flight capabilities or passengers. Have fun with passengers when an automated passenger drone needs to quickly change thrust direction because of wind shear.
That’s a matter of changing the PID compensations, it’s not a difficult problem. Do you think all drones can do a barrel roll? There are nonagile drones in operation right now, in swarm configurations. Not as large as a car, but again, that’s a matter of tweaking the PID, nothing more or less.
To put it another way you don’t restart learning how to drive a car each time you drive a new car. You tweak your internal pids to that of the new car. The difference is humans are stupid, slow, and have terrible coordination that can’t be immediately transferred over a fleet.
There are a whole lot less obstacles and unknowns in the air, as well as more planes (ha!) of separation available than a car.
When flying, you don’t really need to worry much about random pedestrians, for example.
If the entire system were completely automated, from the car all the way to ATC, and it’s essentially a taxi that you just tell what location to go to and it handles the rest… well that’s basically air traffic today minus the automated ATC part. (That isn’t to diminish pilots at all; just that I think it’d be a lot easier, in general, to replace a pilot than a taxi driver with automation. They’re both still extremely challenging problems.)
If you ignore take off and landing, birds, weather conditions, and everything else that makes flying more complex and dangerous than driving on the ground, sure.
And if you ignore construction sites with high cranes and not documented buildings. Or overland high voltag power cables, wind mills, hobby drones, and local variations of birds.
It‘s just taking the complex challenges of autonomous drivinf into the third dimension. Making it even more complex.
Other than the takeoff and landing, cars have to deal with those obstacles as well.
A computer running a citywide automated traffic system for cars would have all the same complexity, without the ability to separate traffic in three dimensions.
So yeah, if you ignore the parts that make it more complicated it seems easier!
If it is windy, it is far more complicated than driving on the road, especially in cities with taller buildings. Like not crashing into buildings is far harder than applying the brakes when there is ever changing wind shear that you can’t see. This applies to most days in most cities.
It isn’t that complex. The problem with current autonomous driving is the car can only infer what other cars are doing and what is around it, especially if we are talking about an autonomous car with idiotic vision mapping without lidar.
With a flying car that is directed by an AI that knows where every other flying object is, what every flying object is going to do, and the locations of every stationary object based on maps and lidar on the vehicle, you can keep collisions far less likely. Taking the human control out of the picture improves the conditions substantially.
I wouldn’t trust a flying car at all, but I would trust an autonomous one with AI ATC far more than an autonomous car going through a construction zone on a highway in a major city during rush hour.
Spoken like someone from a culture where drivers are the only thing around because they have gotten so used to ignoring pedestrians, bicyclists, animals (wild or otherwise),… that might be found on the road and hasn’t considered what else might be in the air at all.
You would have a point if the flying car didn’t have to take into account the wind, updrafts, downdrafts, wind sheer off buildings, and a ton of other flying related stuff that helicopter pilots need to take into account which are barely noticeable to cars the majority of the time.
Then there is landing, the most dangerous part of flying. Imagine if the emergency braking in a csr needed to stop the car without spilling a liquid from an open cup.
Being on the ground is far less complex than flying, otherwise getting a pilot’s license would be easier than a driver’s permit.
This magically perfect AI would work on the ground too. I mean, it knows what birds are going do, why not people?