Paul C. Schutte, NASA Langley Research Center
Many papers and discussions in Aviation Human Factors take the pilot's physical presence on the commercial flight deck as a given. Statements such as, "Until significant advances in automation are made, the human is required..." or "For the foreseeable future, pilots will be necessary ..." are made without much analysis. Once this premise is stated, we often move on to defining appropriate allocation of functions between humans and automation on the flight deck. Most of these allocation decisions address overcoming human limitations and improving situation awareness and workload. However, it seems that these decisions fail to return to the basic question. Why is the human necessary? We need to explicitly address how function allocations and design directly support the requisite role of the human.
The purpose of the panel session is to define why the pilot is necessary in the flight deck. It is not to define the reasons that humans do not perform well at some tasks nor to directly debate their presence. It is hoped that after making these required duties explicit, the discussion can then turn to how best to enable and support the pilot in this requisite role.
Capt. Richard B. (Skeet) Gifford (Ret.) NASA Langley Research Center
Airplanes are much more reliable and therefore much safer than they were when I began my career in aviation. While the pilot does not encounter serious emergencies with the frequency that he once did, vigilance is still required to operate the flight safely. The first airplane I flew with the Air Force, a C-119, had 14 procedures in the Emergency checklist. Those 14 procedures took care of just about anything that would happen to the airplane. In contrast, the B-777 has procedures for over 180 non-normal procedures. Several examples are included which describe incidents where the pilot on the flight deck corrected malfunctions of automation or made other inputs to preserve the safety of the flight. The examples are drawn from my personal experience and from incidents that occurred while I was a Flight Manager. These examples include map shift, uncommanded thrust reduction, spurious calculations by the Flight Management Computer, failure of both Flight Management Computers and database errors.
We must recognize that complacency can be the unintended consequence of reliability and automation. Forty-five years ago, failures were predictable and often catastrophic. Today, failures are infrequent but can be subtle, unpredictable, and difficult to analyze. In automating human error out of the cockpit, we have sometimes merely traded it for human error on the part of the engineer.
Flight crew members should have better training in problem recognition and analysis. Engineers should design systems to keep the pilots in the loop and provide sufficient operational flexibility to enable the pilot to effectively deal with unexpected situations. The NTSB should provide a more sophisticated analysis of accidents so that the industry can correct the root-cause of the failure.
Victor Riley Honeywell Technology Center