The heavy jet has been vectored to intercept the final approach course and slowed to approach speed. Flight attendants have made the cabin announcement to buckle seat belts, stow all tray tables, and return the seat backs to full upright position. Turned down final, the jet descends toward the airport.
From a different airport, another jet has been restricted to a low altitude for arriving traffic but allowed to accelerate to the maximum speed for its altitude. Clear of the immediate airport area, the jet is vectored across the final approach path of the first airport, on toward its planned departure gate.
Two more hand-offs pop onto the controller's screen from another sector, dividing his attention. A readback error from one flight requires two transmissions to ensure that the assigned altitude will be flown.
In a mere seven seconds, the time it takes for the controller's screen to refresh, the passengers on the left side of the arriving jet can be heard screaming, as the departing airliner claws through the final few feet before impact. They come together; the departing jet slams into the other on a perpendicular flight path. From the ground, the sky appears black with the torrential downpour of aluminum confetti, mangled body parts, and complimentary roasted peanuts.
A second before impact, this controller initiated communications to the departing jet to change course immediately. The call came too late.
I feel for the first time the trickle of sweat down my back as the two data blocks flash red, a conflict. Thankfully, this is only a test. No lives have really been lost. The air-traffic drama unfolds only in the binary mind of a computer.
And yet those frantic moments when the mind has recognized the loss of separation, even as the unskilled hands fumble at the keyboard, seem quite real enough, thank you. I contemplate for a moment that, one, controllers area different breed to be able to juggle numbers, directions, and speeds in three dimensions, and, two, I wasn't of that breed.
The test is the airport simulation, a much simplified but nonetheless thoroughly engaging computer model of real-world air traffic control. Just after the loss of separation, I look around the room at the 20 others being tested to see if I'd grunted some nasty expletive. They were too wrapped up in their own three-dimensional hells to notice.
Meanwhile, on the screen, two more handoffs await my acceptance. Shrugging my shoulders to ward off the knotting pain, I bring them under control and resolve to keep the paying passengers in the 12 other airplanes in my purview from fiery grief.
The Federal Aviation Administration is changing its ways. Proof? The new controller screening program now being implemented at the FAA's Mike Monroney Aeronautical Center in Oklahoma City. New procedures have reduced the ante of controllership from nine weeks to five days, all the while making the ordeal much more humane for the applicants. What's more, the agency has reduced the cost and time investment in the screen substantially.
Used to be, if you had your heart set on being a controller, the process would begin with a general government entry test. Those who have taken it and passed call it a relatively simple task, one which requires above-average spatial and recognition skills, but not necessarily the ability to split atoms.
Once through that test, you would be contacted by the FAA and asked to undergo a physical examination and submit to a background check. Providing the investigators find no skeletons in the closet and the doctors nothing unusual in your bones, you would be put on the list for the next step — the screen.
In the past, this screen would consume a full nine weeks of your life and offer you about a fifty-fifty chance of coming out the other side with career potential in ATC. For the half that didn't make it, the task was often to get the old job back or maybe sign up for a truck-driving school. In any case, a tremendous personal investment, in both time and effort, would have to be written off to experience. A few who failed the first time would be allowed to retake the training, but the percentage who elected to were minuscule.
It's different now and seemingly a far better deal for all involved. Although the exact syllabus hasn't been finalized for the course, the basic idea is this: Take the entry test, go to Oklahoma City for five days, and go home and wait. In a week or two, you will find out from the FAA if you have passed the initial screen, and then, depending upon such diverse variables as the region in which you originally signed up and the availability of training courses, you go right to school to learn the tricks of the ATC trade.
Rather than being interactive with the instructors/examiners and an almost completely mechanical process, the new screening program is entirely computer-based. The FAA says this new system is both more accurate and more objective than before. It is also unerringly unforgiving of mistakes.
While the methods of testing have changed, still about 50 percent of those taking the screen fail. The FAA wasn't really looking to pass more applicants per test, but to cut the investment, both of the agency and of those taking the screen.
I'm chanting now. Watching the potential conflicts on the left side of the computer screen emerge with clockwork accuracy, all the while attempting to retain in short-term memory the call sign of an airliner shown to me on the right side of the screen. It is one of the three major tests in the controller screen and one of the most fun.
I flash my gaze to the left side of the screen to see two airplanes' data blocks heading right for each other. Quickly, I look first at the altitude, only 1,000 feet apart — too close. I press the button to signify a conflict.
Then, a small cursor at the bottom of the screen directs my attention to the right half. On this side of the screen, you are tested on your ability to place an airline flight identifier in your own short-term memory and retrieve it without slowing your pace. Here's how it works: You are shown two identifiers separated by a short horizontal line (e.g., UA122 over AA141). If on the previous screen UA122 was below the line, you are supposed to press a button signifying a match. At the same time, you are expected to memorize the identifier under the line for the next go- around.
This exam, called the Static Vector Continuous Memory test, joins two others to form the basis of the controller screen; they are the Time Wall Pattern Recognition test and the Airport Scenario.
The Time Wall, like the SVCM, tests the ability to divide attention while being able to interpret information and correctly deliver the answer. The Time Wall test begins with a small cursor moving across the screen from left to right, where a thin vertical line awaits it. All you have to do is watch the cursor and punch a button when it is on the line. Okay, the catch is that for about half its journey, you can't see the cursor, so you have to keep a clock running in your head to be accurate. And the cursor moves at several different speeds.
What's more, while the cursor is wending across the screen out of view, a series of patterns pops up, and it's your job to correctly answer if they are the same or different. Soon, those in the course figure out that timing each response to the patterns — say, one answer a second — gives something of a metronome and helps make hitting the wall a bit easier. You are scored on both your accuracy with the patterns and your ability to hit or come close to the wall with the cursor.
Through all this, though, the Airport Scenario makes the Continuous Memory and Time Wall tests seem like mere warm-up exercises. Depicted on the computer screen are depicted two airports and four departure gates. Airplanes are depicted by a simplified data block, including an arrow for direction of travel (limited to one of eight headings, labeled 0 through 7), speed (slow, medium, or fast), altitude (levels 1 through 4), and destination (a letter corresponding to one of the four departure gates or one of the two airports). The airports are wind sensitive, in that at the beginning of each scenario, you have to note the landing direction from a symbol at the edge of the screen. Send one in downwind, and you get an operational error. I know.
You control the airplanes at first through voice commands and cursor control via a trackball. Move the cursor, click the button to highlight the target, and tell him what you want. So: Move, click, "Speed, fast," and the response, "Roger, speed fast." The readback even sounds like an airline captain's. Actually, you only use the voice-recognition feature during the early part of the training, when the scenarios aren't all that tough. Later in the testing, you use a template shown to the right of the scenario screen. Move, click on the airplane, click on the command, and listen for the correct readback. Sometimes it doesn't come, and you have to reissue the command; do anything else beforehand, and you get marked down. I know.
The scenarios start innocently enough, but by the last day of practice, they become tough enough to cause me to break a mild sweat. I have trouble with the trackball — this Macintosh user far preferring a mouse to that monstrosity, but I don't have a choice.
But far and away, the most mind-bending part of the Airport Scenario is the need to divide my attention and issue fast, accurate commands. Forget about coming up with "a system," and forget every bit of pilot knowledge; the ATC system works much differently from the inside, and knowing what ought to be the correct outcome and understanding real- world procedures and constraints only slows you down.
Here's an example of that. Most airplanes have a hard time descending from, say, Flight Level 330 to the initial approach altitude in one seven-second sweep of the radar. And most airplanes can't turn around in their own length, also during that seven-second window. But for the sake of testing, the airplanes in the Airport Scenario can, and I have to shake off some of my pilot-think to take advantage of this.
According to the FAA, knowledge of the air traffic system isn't required or even desirable in controller candidates. Only the ability to divide attention, respond accurately, keep a broad mental picture of the airplanes' movements — those basic mental and hand/eye skills are all that really count.
And I find I just don't quite have them. Even on the final day during the morning warmups, I can't seem to expand the envelope. During each of the scenarios, I reach a point where I'm completely overloaded, doing my best just to keep the dozen or so airplanes under my control...well, in control. While some of the others in class are working on getting their delay scores down, I'm working feverishly to keep the airplanes from slamming into each other. It's not easy.
Scoring of the scenario is accomplished on several levels, with perhaps 20 ways to add to your score. (A zero is perfect.) Basically, you have the safety issues, which include keeping the airplanes separated by 5 miles or one altitude level, and the procedural items, like making sure they go out the right departure gate or don't try to land downwind. Then there's the delay score. Each airplane on the screen has a theoretical ideal time from where it enters the screen to its destination. Any delay, like keeping it at a slow speed or excessive vectors, adds to the delay score, as does waiting to accept a hand-off.
So while some in the course were crowing about getting the delay score down to just a few minutes, I was desperately trying to keep the blood off my hands.
The letter from the FAA came a few weeks later, and I was slightly surprised to find the first word in the first paragraph "Congratulations." The score: an unimpressive but solidly in-the-hunt 79; you have to get 70 to pass.
Fellow pilots, worry not. While I might be eligible to go to controller screening, you can rest easy that I won't — in reality, I had not been scored along with the rest to keep from influencing the results of those who really wanted to be controllers.
While I don't want to change seats with the controllers, you can be sure I walked away from the screen with a newfound appreciation of those individuals sitting in front of the radarscope. We as pilots see only the most superficial aspects of the controller's job. I had always suspected this to be the case, but after a strenuous week in the clutches of the FAA's newest testing tool, believe me, I know.
