There’s No Silver Bullet in Effective Loss of Control In-Flight Mitigation

Unlike technological wonders like TAWS or TCAS, there is no single solution to the Loss-of-Control In-Flight problem, which remains today the biggest cloud in that otherwise clear blue sky. Technology, like flight envelope protection, can certainly help significantly. However, regardless of the aircraft, the prevention techniques for preventing upsets (like stalls) remain.

Place an average modern-day pilot in conditions leading up to an airplane upset situation; hindsight tells us that a significant number may not recognize and avoid it, and also may not be able to recover from the resulting upset. While the areas that are statistically less threatening are thoroughly trained, and whereas evidence shows that LOC-I is the number-one threat, it is not trained adequately . Today, we simply do not train  pilots consistently to avoid, prevent and recover from upsets. In fact, some contend that practical test standards for stall training that have emphasized “minimum loss of altitude” may be leading to negative training transfer when it comes to avoiding and recovering from these threats.

So, how do we provide UPRT today, and why is it inadequate? Again, looking back at the incident records, we can characterize that upset prevention and recovery involves three distinct and critical levels of mitigation. First, there is awareness, which includes a thorough understanding of aerodynamics, and knowledge of causes of airplane upsets to further enhance aeronautical decision making. In the event a threat begins to emerge, the pilot must exercise recognition and avoidance in order to stay clear of the threat, or correctly counter a rapidly developing flight attitude or flight envelope excursion without further compounding the situation. However, distractions, failures, or other causes can still lead to an actual upset; hence, the final hope is to exercise an effective recovery to bring the airplane back to a controlled state. In some situations with large transport aircraft, such as the incipient spin, or worse – a fully-developed spin, recovery may not even be possible, making recognition, avoidance and prevention the most critical training results. The conclusion: we need to train pilots through a combination of academics and practical skills development.


For the academics, as one example, there exists today an industry-developed guide book called the Airplane Upset Recovery Training Aid. Ask the average pilot, and chances are high that they will never have heard of it. It is not commonly integrated into training, and its use is not mandated. Furthermore, this aid is limited to large (100-plus passenger) swept-wing jets, while these problems also occur in regional liners and turboprops.

Practical Training in Airplanes and Simulators

Secondly, it is not practical to teach upset recovery or even recognition and avoidance in actual transport aircraft. This is when we turn to the use of flight simulators.However, simulator data are incomplete outside the centre of relatively small, measured envelope. Simulators cannot induce the forces, particularly the vertical g, encountered in upset conditions, nor can they easily create the sudden startle that occurs when a pilot discovers he or she may have very little time to apply life-saving control actions. There even a danger: The NTSB has identified several aircraft upset accidents in which in-appropriate use of simulator, and the in-ability of an instructor to train and evaluate when in-appropriate and even dangerous flight control inputs were made by pilots in training.

Then, we have the aerobatic-capable airplane, or even in-flight simulators, that teach maneuvering skills and exposure to an all-attitude all-envelope environment. Familiarization with the limit load, and given proper instruction, training of this kind can be very effective. The transfer of these skill sets to the multi-crew, glass cockpit, automation-enhanced environment takes place outside the actual in-flight lesson through a process of “differences training”. In this way, flight simulators act as the bridge between basic skills, and application in the target environment.

What can we do in today’s simulators and training curricula to deal with the most significant UPRT issues? Can simulators be effectively and affordably modified to represent an extended envelope? How do we deal with regulatory issues for these data? How accurate does it need to be, and what do we risk if there are differences compared to the real-life situation? Many trade-offs have to be considered.

How Accurate is Good Enough?

Preliminary conclusions were drawn after the ICATEE group flew several cases in the Boeing 737-800 FFS in December 2010 at the FAA Oklahoma City Technical Center. Enhanced data can help pilots in becoming aware that, while managing un-loading during a stall, a dynamically unstable, inadvertent roll can be encountered as the non-linear aerodynamics come into play. Because we have become so reliant on the simulator, current flight simulator data can lead the pilot into thinking that an airplane remains docile and controllable under these conditions. This is clearly not the case, and the regulatory experts in ICATEE also agree that enhancements, albeit representative and not even type-specific, may be a major step forward. Fortunately, with the involvement of research organizations like NASA, UTIAS and several academic experts, we will have the advantage of being able to extend existing models, and carry out more detailed analyses before drawing final conclusions. See the figure below.

Why Train in the Stall Zone?

An obvious question is, why would one ever want to train beyond the stick shaker. Colgan Air, West Caribbean, Pinnacle, West Caribbean and even a near accident of a USAF C-5A all occurred with the stick shaker activated, with plenty of opportunity to recover, and where prevention did not occur in a timely manner. Inattention, inadequate or improper use of automation are also cited as causal factors. Therefore, exposing pilots to the general nature of these conditions can be a worthwhile eye-opener: It’s potentially lot harder than you could have imagined: Don’t remain there, and DEFINITELY avoid ever coming there.

Secondly, in the USA, it has become mandatory to train in the fully-stalled region in flight simulators. Proposed rulemaking changes, including the need to train pilots to avoid and, if not avoided, to recover from stalls, are intended to contribute significantly to reducing aviation accidents. As stated in USA Today (11 May 2011), when referring to the Colgan Air 3407 crash, currently, pilots are trained to avoid entering a stall at all costs but are never shown how to recover once they have entered a stall. “One key to preventing such accidents in the future will be more realistic training in simulators”, according to FAA Administrator Randy Babbitt.

One consideration of ICATEE is that simulators with reasonable “class-specific” (such as “conventional, swept wing, under-wing engine”) model enhancements beyond the currently validated envelopes may be sufficiently suited to provide the necessary training for avoidance and recovery, under the guidance and monitoring of specially-qualified instructors. Enhancements to control force and cueing, using objective testing methods beyond those applied today, may also further enhance the feedback to the trainee on the dynamic characteristics. ICATEE’s investigations will continue.

Surprise and Startle

Hindsight is easy when staring at a YouTube video of a crash replay. However, these accidents do not occur in a vacuum. When suddenly faced with an event, a pilot may become surprised. The result is often what psychologists call “startle”: the pilot’s decision-making capabilities and intelligence can be narrowed down to basic primal instincts, and often, it is in this condition that the rote learned or first-reflex are applied. In some cases, the reaction has proven wrong, and the chances of applying corrective action are minimal.

Can we achieve anything adequately similar in a ground-fixed simulator, in a world where the first pilot to undergo a surprise event posts the findings on Twitter or Facebook to the “benefit” of every following candidate for that training exercise? If this is desirable, at the very least, the instructor needs a significant “bag of tricks” to apply effective distractions. Even then, the required magnitude, quality and relevance of startle/surprise factor, as it relates to UPRT awareness and skill development specifically, cannot likely be fully accomplished through ground-based simulation as a standalone resource in the near term.

Comments are closed.