The Human Factor: Finding the Right Balance

The answer to the debate between the virtues of hand flying and using the autopilot lies somewhere in the middle.

Autopilot

Autopilot

It seems like in many areas people are divided into two camps, with each side strongly supporting its way of thinking while disparaging any other approach. Sometimes the pendulum of opinion swings from one side to the other, never pausing to stop in the middle. A balanced approach is usually more reasonable, applying the appropriate measures at the appropriate times.

One example of this in aviation is the autopilot versus hand-flying controversy. As an instructor at FlightSafety and SimuFlight and a check pilot for the Civil Air Patrol, I have seen both sides of this issue. For some pilots, "gear up, autopilot engaged" seems like one step in the checklist. As soon as they have achieved a positive rate of climb and raised the landing gear, they engage the autopilot, and it stays engaged right up until shortly before landing.

At the other extreme there are pilots who seem to feel that autopilots are for wimps. They hand-fly the airplane right up to the point that they level off at their cruise altitude, and then disengage the autopilot as soon as they initiate their descent for landing. Other pilots seem to think there is a rule about not using the autopilot below 10,000 feet. Some instructors and check pilots “fail” the autopilot at the start of a lesson or check ride and won’t allow its use at any point in the flight.

A Balanced Approach
Right in the middle of these two groups are the pilots who see the autopilot as a tool to be used to relieve the pilot from the duty of manipulating the controls so he can accomplish other tasks such as reviewing checklists, communicating with ATC and preparing for an approach. They recognize that a properly functioning autopilot can fly the airplane as well or better than we can and can free us up for other duties, thus increasing our overall situational awareness and making it less likely that we will make a mistake or miss a call from ATC.

The FARs recognize the advantage of using an autopilot in single-pilot operations. While Part 91 has no regulations governing autopilots in small airplanes, Part 135 allows single-pilot operations only if the airplane has a functioning autopilot. In the same vein, the Mitsubishi MU-2 SFAR requires a functional autopilot to operate single-pilot at night or in instrument conditions. However, just as you can lead a horse to water but can't make it drink, you can require a pilot to have a functioning autopilot, but there are no regulations in Part 91 or 135 that require the pilot to use it.

It is up to each pilot to decide how best to use the autopilot, just like any other tool in his flight case. Because pilots are different, some may want to fly with the autopilot engaged most of the time, while others may prefer to hand-fly the airplane. I have no problem with either approach as long as the pilot can operate smoothly both with and without the autopilot. I get worried when I fail the autopilot and discover that a pilot has been using it as a crutch to cover up a lack of basic piloting ability, or when I ask a pilot who has been hand-flying the airplane to fly a coupled approach and discover the pilot has difficulty using the autopilot.

Every pilot needs to be competent at hand-flying the airplane, and it takes training and practice to maintain an adequate level of competency. However, human factor experts are discovering that the constant autopilot use that is typical in modern glass cockpit airplanes is leading to a decrease in basic hand-flying skills. Recent airline crashes have been traced back to pilots who didn’t have an ability to fly an airplane by basic pitch and power attitude flying. When the system crashed, they crashed.

Flying using the autopilot is a skill that is no less important. It takes a significant amount of training and practice to achieve the ability to use the autopilot smoothly throughout the flight from climb-out to final approach. It can be a thing of beauty almost like a ballet to watch a pilot who is skilled at using the autopilot smoothly shift between modes, activating the right mode at exactly the right moment to achieve the desired result. On the other hand, it can be painful to see a pilot struggling with the autopilot, getting more and more confused as he tries to figure out how to get it to do what he wants. A tool that can free up a pilot and increase his situational awareness can quickly become a deadly distraction from even the basics of flying an airplane.

Accident reports make it clear that just one erroneous mode selection can quickly lead to a crash, and with the increasingly complex interface between sophisticated autopilots and complicated glass cockpits, it is becoming very easy for even an experienced pilot who is not fully trained, current and competent on that system to get into trouble. In one example, a pilot who had 1,300 hours including 400 hours of instrument experience, most of it flying a Beech Duke with conventional gauges, crashed shortly after taking off in a Cirrus SR22. The pilot had taken the VFR Cirrus transition training course but decided not to take the IFR portion of the training. He had flown only 20 hours the preceding year, and only four of those were in instrument conditions.

Due to lack of adequate training, practice and currency, this pilot with lots of experience flying a complicated high-speed multiengine airplane with round-dial instruments was totally unprepared to operate a single-engine airplane with a complicated integrated glass cockpit and multimode autopilot. On his arrival with the weather right at minimums, the pilot missed three coupled ILS approaches because the autopilot had not been properly armed to capture the glideslope before landing on his fourth try.

Instead of realizing that he was not ready to operate the SR22 in low instrument conditions, the pilot took off again into the 200-foot overcast. The avionics data collection system recorded that the pilot engaged the autopilot five seconds after takeoff. He was cleared to fly runway heading and climb to 3,000 feet, and the autopilot altitude preselect mode was properly programmed to climb to an altitude of 3,000 feet at 850 fpm. However, instead of activating altitude preselect, the pilot engaged altitude hold, causing the autopilot to immediately level the airplane and attempt to maintain 940 feet msl, which was only 61 feet above the field elevation. The airplane went through a series of increasingly wild gyrations as the pilot attempted to regain control before crashing in a partially inverted attitude just over four minutes later.

Each pilot should strive to achieve a balanced approach. On the one hand, we need to maintain an ability to hand-fly the airplane at all airspeeds and in all configurations. This will likely not be a problem for a pilot who flies only local VFR flights, especially if he likes to challenge himself by doing basic maneuvers like slow flight and steep turns. However, pilots who fly most of the time with the autopilot engaged, and those who don’t have an autopilot but spend most of their flight time in level cruise flight, need to regularly go out for a review of basic airwork skills so that the next time they have to slow down behind an airplane in the pattern or go around just before touching down, they are up to the task.

At the same time, pilots operating airplanes equipped with autopilots need to achieve and maintain the ability to smoothly use the autopilot to its best advantage in all phases of flight before they attempt to use that auto­pilot in actual instrument conditions. For the new integrated glass cockpits this will require a fairly extensive training course with plenty of practice in the airplane on different types of approaches and simulating different possible situations. This needs to be followed up with consistent use whenever possible. Some of these new systems are complicated enough that if you don’t use them for a few months you will need a refresher session before attempting a flight in instrument conditions.

It is up to each of us to assess our own situation and determine what is necessary to maintain a safe level of basic flying and autopilot skills. One of the main reasons I took myself off flight duty for several years was that I had been flying a Cessna 182 with a Garmin G1000 for the CAP, and I found I simply was not flying it enough to maintain what I considered a safe level of competency. I have no problem physically flying an airplane after a long period of inactivity, but even a couple of months off left me scratching my head and feeling totally inadequate at correctly programming and using the G1000. While I was legally current and qualified to fly the airplane, my lack of currency with the G1000 would have distracted me from other critical functions. That was a risk I was not willing to take.