# Designing a traffic system for flying cars

Well folks, 2015 has come and gone, and we still don't have widely-available flying cars. While there is some promising work being done on "flying cars", which are more like road-capable airplanes, it looks like we're a long way away from the type of flying cars I'm imagining from pop sci-fi. Thus, I'll abstain from using the "near-future" tag on this one.

Okay, let's assume we're in a future Earth where flying cars are as ubiquitous as ground automobiles are today. Their price and the cost of operating them are comparable to today's autos as well. Also, flying cars are basically hover cars, and can be stationary in the air, or move fully in three dimensions, as you might expect a flying car to be able to do. To clarify, the bottom of the flying car must always be pointed down (unless you're pulling off some crazy car-chase maneuvers), and it can float up and down along a z-axis, move foward and backward as normal along a y-axis, or move left and right along an x-axis freely, without having to turn to point the front of the vehicle in that direction. However, it would move along the x-axis slowly, and do so by "banking" the vehicle. Also, while stationary, it can rotate to point the front of the car in any direction along a two-dimensional plane parallel to Earth's surface. The front of the vehicle can also tilt to an incline or decline of a limited amount, let's say by 25 degrees. And, for whatever reason, none of these flying cars are "self-driving".

In some fictional futuristic worlds that contain these types of flying cars, you often will see traffic signs, signals, and patterns that hail back to modern-day Earth. You might see a normal grid-pattern of cars in a city, but the grids stack up along a z-axis, and at each intersection, a stack of floating common traffic lights. You might even see a floating highway - a literal highway - that has floating lane markers and exit signs.

This type of traffic management system has always seemed so contrived to me. There has to be a better way, especially considering that vehicles can move in three dimensions (or four if you're a specific DeLorean). However, I'm struggling with envisioning that better way. I suppose a future traffic system for flying cars would incorporate some degree of what a ground traffic system would look like today, with some of what today's air traffic control systems use as well. How could air traffic principles and ground traffic principles be combined in order to produce an effective traffic system for flying cars?

• Why would there be any "Traffic" could every intersection not become a Cloverleaf intersection suspended above the ground? Jan 5, 2016 at 22:28
• “none of these flying cars are "self-driving"” – realistically, all of them are. Jan 5, 2016 at 22:39
• Yes, the idea of flying cars is terrifying unless the average citizen doesn't have any control whatsoever over the vehicle, aside from inputting a destination. I expect them to get the destination part wrong a double digit percentage of the time too. Insurance companies would never allow it either. Jan 5, 2016 at 23:00
• I don't think this deserves a -1 rating, it's not that bad of a question Jan 5, 2016 at 23:15
• Possible duplicate of Urban Planning in 3-Dimensions Jan 6, 2016 at 21:01

Real pilot here.

Here are the air traffic management rules for what I shall call "Class V Airspace" :

1. Class V airspace generally only exists in the most congested urban areas of the United States, where normal air traffic procedures cannot sustain the high volume of VTOL traffic. Class V airspace generally consists of the region between 1000 and 3000 feet AGL. In more congested urban areas, Class V can exist between 1000 and 5000 feel AGL, with higher limits for airspeed in the region between 3000 and 5000 feet. Air Traffic above Class V airspace shall adhere to VFR or IFR flight rules, whatever that airspace is classified as, whether it be Class B, C, D or E.
2. Air Traffic inside of Class V airspace must travel between 100 and 150 knots, with altitude mapped to aircraft magnetic track such that at odd levels of 1000 feet (For example: at 1000, 3000, and 5000 feet) the aircraft must be traveling North (0 degrees). As the aircraft climbs (or descends), it must adjust its magnetic ground track heading to map to the altitude it is at, so that as it makes an ascending right-hand turn (or descending left-hand turn), its altitude is an even multiple of 1000 feet at the instant it is heading South (180 degrees). This allows the aircraft to make a standard 3-degree-per-second climbing turn at 1000 feet-per-minute while only conflicting with other aircraft either directly in-front or behind it. Aircraft must yield to conflicting aircraft in-front by passing on the right.
3. Air Traffic below 1000 feet and above 700 feet shall not exceed 100 knots, and shall yield right-of-way to any other aircraft on its right.
4. Air Traffic above 400 feet and below 700 feet AGL shall not exceed 50 knots.
5. Air Traffic below 400 feet shall not exceed 25 knots.
6. Landing aircraft have right-of-way over landed aircraft at vertipads.
7. In Class V airspace, separation is maintained using visual see-and-avoid techniques and ADS-B telemetry. Conflicts are handled using standard aeronautical right-of-way procedures.

And there you go!

It would probably take the rest of my night to provide the calculus of why this works and requires nothing more. In short, having access to the volume of air above even a large city far exceeds the traffic density that would have to exist to require a more complex management solution. Traffic congestion with automobiles is a result of their 1.5 dimensional nature: In a car, you can only move forward and possibly switch lanes. We like this as drivers because it make all the possible collision vectors collapse to only a few possible directions. Modern aircraft traffic rules somewhat also do the same thing. (See this link and this link).

By mapping altitude to heading, we collapse the possible collision vectors to only ahead and behind. This method would work great for the average urban area. In areas where traffic density is much higher, it would make sense to have "climb" zones and "descent" zones where the mapping is still followed, but in addition only standard climbing or descending turns are allowed. Otherwise, aircraft are free to fly straight lines between points; getting to your desired heading only requires that you make a climbing/descending circle until your pointing where you want to go.

The altitude mapping method works great for vehicles trying to get around, but how do we handle take-off and landing? This requires that we remove the mapping requirement for the regions where take-off and landing operations happen. To resolve the collision hazard, we restrict the cruising speed considerably. By restricting flying speed in the lower layer near the ground, any conflicts that could result in a collision will happen slow enough that either one or both pilots can react with ample room. The worst-case approach speed is 50 knots for a head-on situation below 400 feet, and 100 knots for the 400 to 700 foot region.

Transitioning to the more busy and fast layer from the lower layers requires starting on a north heading as your aircraft passes through 1000 feet AGL. Aircraft climbing to this level will tend to align to north as they approach 1000 feet, at which point they must be traveling at least 100 knots and at most 150 knots. Thus, any conflicts are happening at 50 knots in-front or behind.

In the case of aircraft transitioning to the slow layer from 1000 to 400 feet, the aircraft will deviate from North to various random headings. The possible conflict vectors expand to a field of regard of 360 degrees (a conflict can come from any direction), but by the time these aircraft have descended to 400 feet, they should only be traveling 25 knots, which again yields conflict velocities of only 50 knots.

The most important feature of this approach is that the occupant(s) of the vehicle are in control, and have full proficiency and ability to make decisions based on what their instruments and eyes tell them. I've always been severely bothered by the various inventors and futurists that think that flying ought to be only handled by automated systems, with the occupants at their mercy. Flying is a talent and a freedom that the average person can master and should be able enjoy.

• This answer is so good. I got my Ground School book for Christmas and was reading up on Class A, B, C etc space....your answer is just brilliant. Jan 6, 2016 at 16:00
• Great answer, but I have to disagree about the automation aspect. Driving is a talent and a freedom that the average person can master as well, but people are still terrible at it and unnecessarily kill tens of thousands each year. Jun 14, 2018 at 17:33
• So, this would require every car to be ADS-B In/Out (not just ADS-B Out) equipped? Or should we just skip the preliminaries and go straight to automatically executed TCAS RAs?
– user
Jan 16, 2019 at 19:23
• @aCVn Assuming VFR conditions, this system would only require accurate altitude and heading, so as to collapse all collision vectors to ahead-behind, at which point it is basically see-and-avoid that works. ADS-B (Out) would be a supplement, not a hard requirement. Losing ADS-B does not totally break the system, in the same way that drivers treat an intersection as a 4-way stop when the intersection traffic control lights are disabled. Now, in IFR conditions, ADS-B would be an absolute requirement. But at that point, everything would be on a flight plan anyway. Jan 18, 2019 at 4:58

I made a computer program in my architecture university days to test this very hypothesis.

As I discovered, the best system is a very simple one: your direction is a direct linear relationship to your altitude.

I created a random series of tall buildings, and populated the whole sky at all altitudes with flying cars. After testing convoluted collision detection algorithms, 'streams', blocks, intersections and others, I have discovered that the most simplest solution is that as you ascend, you also turn clockwise. Everyone has the same direction at each altitude.

It was amazing and wonderful to behold, because:

• there were no collisions - at any given height, everyone goes the same direction.
• all you need to watch is speed and who is in front of you - just like we do now on a freeway.
• if you need to climb, you rotate clockwise as you move forward
• if you need to descend, you rotate anticlockwise as you move forward
• you can find your way to any point in space, by ascending to the right altitude, then making a 'b-line' to the point, then spiralling to the right altitude. Every point in the sky is accessible by you.
• for obstacles (like a building) the flow will go around the obstacle, even this presents no crashes. For instance, if your path runs into a skyscraper, everyone descends or ascends (whichever the easiest) to change their direction around the obstacle, still within the rule, then ascends or descends when past the obstacle, following the original path. Easy. With a thousand cars, they all flowed around the buildings like water, with not a single collision.
• NO TRAFFIC RULES, ie. no complicated things like 'give way to your right' or 'stop at intersections' or even 'look over your shoulder'. Simply follow the rule that your direction rotates with your height.
• NO COMPLICATED TRAFFIC SYSTEM, no centralised authority needed, no need for communication between cars
• it was crude, dumb and simple - just the thing that people can understand and is completely infallible.

I cried out in joy after discovering the solution and thought 'yes this could work, if only someone could invent a flying car!'.

• I've been wanting to try simulating this for a while (you and I both came to the same conclusion), is your code online anywhere? Apr 13, 2018 at 1:31
• Please do publish a paper or something. Your discovery might be useful for the future, in case flying cars do become invented. Knowledge is lost unless it is published and reliably preserved for centuries. Nov 14, 2019 at 21:02
• how is it possible that no crashes occur at the point of changing altitude? when at the same time one car is going up and one is going down, it is not excluded that their paths do not cross Dec 3, 2019 at 12:30
• @Zavael the chances are very rare as not only would x and y axis need to be the same but the altitude level would need to be exactly positioned - however this possibility may be negated by a proximity alert and solved by one of the cars, perhaps the ascending one, slowing slightly to allow the other to descend. This would only happen in free space as around buildings it wouldn't happen because cars would 'spiral' around them and would not intersect. Galaxy and Steve I'll see if I can get my code up.
– flox
Dec 3, 2019 at 13:52
• @flox in my opinion, mainly in cities is the chance pretty high. In cities there are many point of interests where cars are going to arrive from many directions and depart to many directions. in cities (comparing to highway), cars are more prone to change directions to find the shortest (or fastest with lower traffic) paths. So without intelligent car, humans would be unable to prevent collisions, its not totally collision free by itself. But it is an interesting idea indeed! Dec 3, 2019 at 14:14

The best plan would not consist of highways, intersections, lights, signs, etc. Instead, a flight control system (like the FAA) would be the ideal system. Even assuming the cars are driven by humans, you could input the destination into the car, which then uses computers to log a flight plan. You then follow the flight plan.

The real signals would be destination signs, Lets be honest - Walmart and Victoria's Secret look the same from above, but you goto them for different reasons. Parking and take-offs will need signals, indicating number of parking spaces available, what floor you can land on, etc. That's where the complexity will really lie. The open highway will be truly open, except in the mind of a computer.

• A common preconception people have about the National Airspace System (aka a "flight control system like the FAA") is that planes are required to file flight plans. This is only true for IFR flights, which, among other things, allow you to fly in Class A airspace (The really fast layer from 18000 feet and up). A large proportion of general aviation pilots rarely file flight plans, especially when just flying around the local area, as a typical aerial commuter would do. Jan 6, 2016 at 5:29
• @Steve, appreciate your input as a real pilot (I gave you a +1 for your answer). However, I figured the flight plan for local flights to keep the tens of thousands of cars in the city from intersecting. That said, I figured it would be automated in the land of the future so the flight plan is really just a data file on a database that has been cross checked against the other files. Your additional recommendations seem the best course for default flight patterns.
– Mark
Jan 6, 2016 at 8:06