ATC is Not the Problem!

ATC is Not the Problem (Managing the Skies, Spring 2022)

For the last 40 years, Eurocontrol, FAA, NASA and others have all worked extremely hard and spent $100s of Billions to fix delays, congestion and reduce CO2, but the benefits have eluded us.

During these 4 decades that I have been involved with ATC and airline operational issue, the ATC proposed solution has always been 10 years and $100 Billion into the future.

• This was the case in the 1980s with the Microwave Landing System (MLS), curved approaches, Advanced Automation Systems (AAS), Initial Sector Suite System (ISSS), etc.;
• The case in the 1990s with the Future Air Navigation System (FANS), Global Positioning System (GPS), and FreeFlight Systems
• The case in the 2000s with Controller Pilot Data Link Communications (CPDLC), Standard Terminal Automation Replacement System (STARS), Required Navigation Performance (RNP), Automatic Dependent Surveillance — Broadcast (ADS-B), and Automatic Dependent Surveillance — Contract (ADS-C)
• The Case in the 2010s with NextGen, SESAR, ERAM, and enroute climb — and now, still is the case in the 2020s with NextGen/SESAR

Many argued that the reason is that we have not spent enough money. Yet ANSPs (Air Navigation Service Provider, government aviation authorities and Air Traffic Control systems) and airlines have literally spent $100s of Billions. And again, did I mention the benefits have eluded us?

Other have argued that, as mostly government agencies, ASNPs are too bureaucratic to accomplish such a large task. Yet countries where the ANSP is not a direct government agency, delays and congestion still exist. Still others have said ANSPS don't have the resources or talent to solve this.

Yet, I find it difficult to believe that the expertise within the FAA, NASA, Eurocontrol, Air Services Australia, etc., are less than needed if delays were actually an ATC problem. They are not

Finally, others have argued that the problem is too large to fix, that weather is too unpredictable, that airline schedules cause the problem, we need more runways, or..., or,... or. But you get the picture, lots of reasons why it won't work, and still no solution.

So in the end, most focus on the most visible participant - ATC as the culprit, and ATC is in control around the airport where most of the delays are visible and where most of the effort and money are funneled.

But let me be clear, Air Traffic Control (ATC) is NOT the problem; nor is it the solution.

So how can such a long held belief not be true? The belief that only ATC can fix this stems from the fact that 40 years ago, when the above assumptions were made, we did not have reliable weather forecasts, good aircraft position data, the necessary communication capabilities or the computational power to drive to look into the movement of the aircraft in real time or the necessary real time predictive analytics. Further, ATC was the only entity that had even close to a wide enough view of the problem to actually do anything, so everyone concluded that ATC was the problem and only ATC could fix the problem.

But time moves on, technology gets better, and since the 1990s, the forecasts, aircraft position data, communications and computational power required are readily available to see what is happening within the world's airspace in real time, and more importantly, with predictive analytics, what will happen "day of", hours into the future.

So why can't we fix this? The answer is simpler than everyone would believe - for the past 40 years, we have been trying to solve the wrong problem.

Variance

So if ATC is not the problem, what is? The answer is variance.

Of course, variance in the movement and flow of aircraft has never been fully considered. But, as the slide shows, the variance in the movement of an airline's aircraft is huge. But instead of removing this completely unnecessary variance, all of the energy, money and focus has been on ATC and technology. No one went back to do the basic engineering, i.e., determine the root cause of the problem before you work on the solution.

Also, let's look at what we now call ATC delays. I believe that actual ATC delays are close to zero. Of course, an ATC equipment failure would be classified as an ATC delay, but these are few and far between.

Next, consider that ATC is a reactive process. If the airlines throw 150 aircraft at Chicago in an hour, when Chicago can handle 120 aircraft an hour, or more realistically, 60 aircraft in half an hour, only one thing can happen - delays. ATC must vector the extra aircraft farther out over Lake Michigan into an extended downwind and final, and ATC takes the blame for the delay. In manufacturing, this is called work in process inventory, which all industries, including the airlines, recognize as very expensive.

Therefore, is the fault of delays and congestion attributable to ATC who can't handle the bunching in the last 30 to 40 minutes (i.e., overload), and are forced to space out the arrivals safely? Or are the delays the fault the airlines who allowed to bunching to manifest in the first place? I choose the latter.

Further, as the slide shows, Time in Queue grows exponentially as a process approaches capacity. During the overloads, we are at or near capacity. As shown time and time again, in numerous other industries, actual wait times (i.e., delays) grow exponentially as the variation increases. This is fact, not theory.

The question then becomes what is the easiest way to resolve this bunching. ATC could try to safely pack them in tighter, a solution they have been working on for the last 40 years with little success, i.e., Defect Correction, waiting until after the problem has developed to correct it. Or airlines can choose the simpler, quicker and much less expensive solution of not sending 150 aircraft into Chicago airspace in an hour (i.e., Defect Prevention).

Now here, you are probable thinking schedule changes, but, unless the schedule changes are draconian, they make little difference. Think 2002 after Sep 11th, when we still had delays.

No, the solution is much simpler and less costly - manage the aircraft arrival flow to remove the variance.

Airlines can quickly do this now by applying business based time flows to speed manage the enroute aircraft to not overload Chicago. By speeding up late aircraft or ones with available gates, and slowing down early aircraft or those without gates the airlines could spread out the arrival flow, both forward and backward in time from a business perspective, so that you have the right 120 aircraft entering the Chicago airspace in an hour, and more importantly, sequenced such that 2 aircraft per minute are time sequenced to the end of the runway.

So instead of randomly overloading ATC, and bunching the flow so ATC is forced to act, and increase time in queue, airlines could manage the arrival flow to provide ATC the right amount of aircraft, at the right time at the right place. Right part, right time, right place.

As proof of how this can be accomplished, consider that in 1995, by simply adjusting enroute speed, United calculated that my on time performance was over 5% better than the average B737 Captain, while using 100 pounds less fuel per flight. Add in coordination across all flights and, as Embry-Riddle concluded, "benefits improve with more flights optimized and complied".

Further, nothing I propose prevents ATC from safely packing the aircraft in tighter. When and if, a big if, this happens, the time based flows would automatically adapt to the new spacing and capacity. In other words, airlines get the best of both worlds.

Now for another contrarian perspective - airports are NOT full. Point overloads - absolutely, full no. Every airport that we have analyzed has shown available capacity, but it is mostly forward in time, and therefore these available landing slots are wasted in today's operation.

To prove this, simply take a good weather day aircraft arrival flow for a full day in the actual order of their arrival and then tighten the flow on paper (well, actually in a computer) to current minimum spacing. Next do the same on a bad weather day. If every slot is not filled, which it isn't, the airport is not full.

The prime cause of delays, congestion and overloads is, as I said above, variance, not ATC, and something airlines or ANSPs (FAA, Eurocontrol, etc.) have yet to fully consider. For example, while analyzing Frankfurt, we noticed delays continued even after the demand dropped below capacity. As you can see in the slide, a variant flow, even at the same demand will cause delays and congestion to add in spacing turns. Now instead of 2 aircraft and one downwind, add in 8 aircraft coming from different cornerposts, 2 downwinds and a few straight-ins, and the delays explodes.

Next, consider an analysis of JFK, which showed that all delays could be eliminated by simply changing the arrival time of the aircraft by an average of 2 minutes, with the maximum required arrival time change of less than 7 minutes. These types of coordinated arrival time changes are all well within the capabilities of any aircraft, on all but the shortest flights.

And the shortest flights could be managed with a coordinated departure time. Not a GDP (FAA) or CFMU (Eurocontrol) time, but a real time, fully coordinated departure time into a fully coordinated arrival slot.

The benefit of this approach is that it does not require airlines to change their schedules, easily crosses FIR/sector boundaries (a hugely important factor), uses the navigation and communication equipment already paid for and in place, is easily coordinated with ATC and can be rapidly implemented within 3 years across the entire US or Europe.

Also, consider what a NextGen/Sesar solution would look like. My understanding of NextGen/Sesar is that ATC will take full control of the movement of the aircraft. ATC would tell the aircraft what time to depart, what time to be at different fixes and what time to land. And these times would be decided with little to no business input by the airline.

So if the gate is occupied and ATC says to land early, you land early and wait for a gate. If the ramp becomes overloaded with aircraft waiting for gates, ATC would slow arrivals. Or if the aircraft had a maintenance problem and wants to land early to provide more time for maintenance, or if the aircraft had a 35 minute turn time and a full aircraft in and out and the airline might want to land early to improve the chances the next flight would depart on time, or if the airline wants to land early because of crew legality issues there would be no easy way to easily communicate this to the ATC solution.

In what business world is having an outside entity, with no interest in the business needs of the users, a viable operational model?

Of course, airlines don't even consider these business decisions in real time today and are happy with their 1950s, fire and forget, wing and a prayer day of operation, where airlines send a Billion dollars of aircraft out of the wing and pray that it comes out Ok. As history and reams of data shows, it rarely does.

And of course there is CO2. Today, we look to very expensive, yet to be commercially viable sustainable fuels, which are at least 10 years into the future. Conversely, above and beyond any current airline sustainability or ATC program, airlines can inexpensively and internally cut their Carbon Footprint an additional 5% by 2025, while also improving profits and reducing delays/congestion. And, again, nothing I propose prevents airlines from using sustainable fuels. When and if, a big if, this happens, the time based flows would still lower the amount of sustainable fuels burned.

As I have talked about for over 25 years, since I introduced FreeFlight back in 1995, airline managed, time based flows are the only solution that rapidly reduces CO2, delays/congestion/costs and solves each and every one of the problems mentioned above.

So we can continue to choose the 10 year, $100 Billion ATC plan, which has yet to provide the desired results. Or we can choose the 3 year, $25 million airline self-help plan which reduces CO2, delays, congestion and costs that FAA and Embry-Riddle jointly proved (FAA Task J Report - Steve Bradford, Dr. Vitaly Guzhva and Dr. Ahmed Abdelghany), Delta Airlines proved (Delta Checklist Publication), Georgia Tech proved (Dr. John-Paul Clark) and GE Aviation proved (Dubai FLOW Report) in actual operations.

Your choice!

Finally, let's recap the benefits of "day of", real time, business based, cloud driven, aircraft time flow management.

1. More passengers where promised, when promised - check.
2. Higher product quality - check.
3. Happier, more satisfied customers - check.
4. Low cost solution, rapid implementation - check.
5. Easily crosses FIR/sector boundaries - check.
6. Less CO2 generated - check.
7. Less noise around airports - check.
8. Less low altitude maneuvering - check.
9. Less fuel wasted - check.
10. Less flight time per leg - check.
11. Higher airline productivity - check.
12. Less dissatisfied passengers yelling at agents - check.
13. Less disruptions and schedule deviations - check.
14. Less flight crew deviations - check.
15. Less costs, higher profits - check.
16. Happier shareholders - check.
17. Less ATC complexity - check.
18. Less government spending - check.

What's not to love?

Real World Business Managed Flow Examples

1.????I would often ask my copilots what time they want to land, something they never really consider, which is amazing to me since airlines sell time. After a short pause, they would answer on time, which is a good first answer. But what if the inbound and outbound are full with a minimum turn time? The airline might want to speed up the aircraft to land 10 minutes early to assure the next departure is on time. Or what if the gate is occupied for 10 minutes after scheduled arrival? The airline might want to slow the aircraft enroute, save fuel, release the earlier landing slot, not congest the ramp or anger the pax who see empty gates, just not theirs. Or what if the aircraft requires a 1 hour maintenance action with only a 40 minute turn time? The airline might want to speed up the aircraft to land 20 minutes early, to allow the next departure to be on time. So the correct answer to what time should the aircraft land is "it depends", something only the airline can decide from a business/system perspective since most changes will impact another flight (not ATC or the individual pilot).

2.????First, let’s look at a flight which lands into San Francisco 30 minutes late for fog. Now of course, this is one of those times when weather is the first tier cause of the delay. But because of the airlines crew scheduling practices, they schedule the aircraft, pilots and flight attendants separately, so when the aircraft lands into San Francisco, the pilots go one way, the flight attendants go another, while the aircraft sits and waits for pilots and flight attendants from 2 other flights.

Now since the crews have to switch aircraft, which means that by the time the new crew arrives at the aircraft, does the aircraft preflight and is ready to depart, they end up off the gate 45 minutes late. And then the flight is planned to Los Angeles at normal speed and altitude, since each flight is treated separately. But given the pressure to reduce fuel cost, yet another cost center, the crew slows down to save fuel, so they land 49 minutes late. But since they were later than the Los Angeles ramp people expected, no one was there to park the aircraft, so they get to the gate 52 minutes late. The same process repeats itself on the flight back to San Francisco, and the aircraft is now 1 hour and 4 minutes late, and then again on the next flight to Denver, which means the aircraft is now 1 hour and 19 minutes late, and on, and on, throughout the day.

Conversely, if the airline kept the crews together with the aircraft, and develop a ‘fast turn’ process to service the aircraft at the gate, such that the 30 minute late arrival for fog departs SFO only 25 minutes late. Also, the flight plan could be calculated to give the pilots extra fuel to allow flying low and fast (best groundspeed), so that it arrives into Los Angeles only 21 minutes late. The ramp could always, yes always, park the aircraft when it arrives, the agent would always, yes always, immediately put the jetway up to the aircraft and open the door, do the ‘fast turn’, and depart only 18 minutes late. Low and fast again, and now it is 14 minutes late into San Francisco. Repeat, and the flight is on time into Denver.

So, instead of 3 legs and the aircraft is over an hour late, it could be 3 legs and the aircraft is back on time. But this takes a system and defect prevention view, something airlines do not understand.

3.????Consider 2 aircraft at the front of a tightly packed arrival queue of 30 aircraft. By identifying/speeding up the first 2 aircraft, moving them forward 2 minutes, the entire arrival queue moves forward. In other words, moving 2 aircraft forward at the front end of a large arrival queue doesn't just save 2 minutes, but saves 2 minutes for every aircraft in the queue behind the first 2 flights, as the entire queue moves forward. This creates what Dr. Clark of Georgia Tech labeled the “draft effect”, thus dropping 60 minutes of flight time and delay from this one arrival queue alone.

4.????Another example is my flight from Portland, OR (PDX) to Chicago (ORD). That day, the tailwinds were in excess of 180 knots, which would put me into ORD 30 to 40 minutes early. Of course, the PDX agents wanted to shut the door 10 minutes early and "push" the aircraft to ORD, since everyone was on board the aircraft (local goal of “shutting the door” to meet D0, or on time zero departure), which I prevented, and we left on time. Next I taxied very slowly, and cruised at a low speed for better fuel mileage, to the point that ATC asked why I was flying so slowly. When I arrived into ORD, I landed 16 minutes prior to schedule, instead of 30 to 40 minutes like all of the other arriving aircraft which were "pushed" off their departure gates to meet D0 and wasted fuel going normal speed.

Of course, when so many aircraft land 30 to 40 minutes early at a hub airport, the gates are still full from the previous arrival bank. This forces ATC to temporarily park and manage aircraft anywhere they can, to the point that as I exited the runway, I couldn't talk with ATC as they were completely overwhelmed with D0 "pushed" aircraft everywhere waiting for their gate. After a few minutes, I was able to break in on the radio, and received clearance to my gate, which was open. As I entered the alley, yes my gate was open, but it was blocked by 5 other aircraft that had just left their gates, which were awaiting taxi clearance to depart.

The end result was that ORD devolved into a classic grid lock situation between the departures and D0 forced early arrivals, as the ATC system and airport were completely overwhelmed. I sat for 20 minutes looking at my empty gate 200 yards ahead, but couldn't get to it. Of course, like everyone else who landed 30 to 40 minutes early, I was late to the gate (20 minutes), even though I landed 16 minutes early.

Could ATC and the airport handled this better, of course? But the real solution was for the airlines to manage their departures by "pulling" the right aircraft from their departure gate so as to not overload the ORD ATC system or airport in the first place. Clearly, if I, as a simple line pilot recognized the problem hours prior (accurate ETA information hours in advance), an airline should have done the same, and prevented the problem from developing in the first place (ala W. Edwards Deming).

5.????Or we often hear the airline delay and congestion problem expressed in terms of the printed schedule, i.e., "You can't schedule 10 aircraft to land at 8 AM and expect everyone to be on time". Of course, if all 10 aircraft showed up at exactly 8 AM, this would be true. The answer to this riddle is twofold.

First, airlines deliver upwards of 80% of their aircraft off schedule (early/late), so the potential of actually having all 10 aircraft arrive at 8 AM is very low.

But the real answer of how to schedule 10 aircraft to land at 8 AM and assure that all 10 are on schedule is for the airline to tactically manage the aircraft so the first one lands at 7:51 AM (assuming a 60/hr. arrival rate), the second at 7:52, the third at 7:53, etc. To do this requires a level of tactical, real time control that airlines currently have all of the data necessary to accomplish, but choose not to do.

6.????Next, let's look at airport capacity. No airport I have ever landed at is "full", as there is plenty of capacity available, even at Chicago before their multi-Billion dollar runway revamp, but it is mostly forward in time. Of course airports are over capacity at certain times of the day (even Boise is overcapacity when 2 aircraft want to land at the same time), but this doesn't preclude reducing delays, congestion and realizing a much improved on time arrival performance.

We simplify the arrival flow problem by looking at the airport as a single entity, i.e., a box. If the box can hold 100 aircraft per hour, allowing a rate of 130 aircraft per hour (or 65 in 30 minutes) to enter the "box" assures a very expensive 30 NM final. Control entry to the "box" and you mitigate much of the current arrival inefficiency and costs (block time, fuel, crew time, noise pollution, etc.).

To make this a reality, the solution is that instead of waiting for ATC to de peak the actual arrival flow backward in time (at around 200 NM from landing), an airline could proactively pull the “right” aircraft off the front end of the actual arrival queue (at 500 to 1,000 NM from landing, or more). By speeding the “right” aircraft at the front of the arrival queue, moving the aircraft forward a couple of minutes, the entire arrival queue moves forward, thus providing the airport a consistent, manageable flow, which would increase airport throughput and improve reliability. Further, an airline could speed up late aircraft, aircraft with maintenance issues or crew legality issues and slow early aircraft or aircraft without a gate.

7.????Finally, why fly fast enroute if your gate is not available. Not only does this waste fuel enroute, it congests the arrival fix and delays other aircraft, takes up a valuable landing slot which should be used by a late aircraft, congests the ramp, and, as proven by ATH Group, leads to increased taxi times while early flights wait for their gate. Further, the airline has ramp workers, fuelers and other secondary processes “standing by”, wasting time and costing money. One action produces lower quality with numerous highly variant and costly effects.

Nothing academic here, just well understood Supply Chain and defect prevention tools from a system perspective.

Given the above, one would think that airlines would jump at the chance to internally implement an FAA proven, independently validated, inexpensive solution that, within months, can improve on time performance, product quality, profits and ATC, while cutting costs, fuel, CO2, noise and daily defects, all with an ROI measured in months, if not weeks.

xxxxxxxxxxxxxxxxxxxxxxxxx`

Additional articles outlining why a Defect Prevention, Operational Excellence solution (>%5 CO2 reduction, >85% A0, <3% day to day A0 Standard Deviation, >8 minute scheduled block/gate time reduction per flight), driven by GreenLandings?, is the path forward to make airlines dramatically better and more profitable. Like Toyota did in the 1980s for the auto industry, all it takes is one airline and/or ANSP to lead the way.

• ?ATC is Not the Problem (Managing the Skies, Spring 2022)
• GreenLandings? Heathrow Interview by Harold Goodwin - Responsible Tourism (video - 46:46, 2020-12-30)
• Aviation Needs a New Direction - Driven by Vision and Leadership (Managing the Skies, Nov/Dec 2019)
• Fastest Airlines in the U.S. (Forbes.com, 2019-06-17)
• Checklist publication - Confessions of an GreenLandings? Doubter (Delta Checklist, 2007-09-01)
• Air Traffic Control Is Not The Real Cause Of Airline Delays (Forbes.com, 2017-03-23)
• Not Working! (ATCA Tech Symposium, Atlantic City, 2018-05-16)
• Parked Planes Cost Airlines Billions (Forbes.com, 2017-08-15)