Cockpit Automation / The Impact

¶ … cockpit automation / the impact of very light jets on FBOs

Cockpit-FBO's

Problems with cockpit automation / the impact of very light jets on FBOs

Paper-1 Problems with cockpit automation

This paper analyzes the pitfalls of automation within the cockpit. Today automation has become more widespread than ever before, especially within the aviation industry, and automation as such is more often than not being blamed for causing great harm, although inadvertently, by increasing the chances for human error, especially when the human being starts to depend on the computer to solve his problems for him. Several experiments have been conducted to find out the depth of this problem of automation and its advantages and disadvantages and to find out whether automation may be worth its while or not. To date, however, it is not clear whether automation carries with it more benefits, or more problems, and this paper helps to understand differing opinions on the same: is cockpit automation good today, or are the inherent problems in automation overtaking the underlying benefits?

Introduction:

Automation pervades almost all spheres of life today; from the small farmer who may grow his daily produce, to the aviation industry that may fly its passengers form one corner of the world to another. The problem is whether automation can be considered beneficial for the aviation industry, especially within the cockpit, or must it considered to be redundant? Do the advantages of automation outweigh the disadvantages of automation? Would it be a good idea to depend more on a human being, rather than on autopilot? This paper will analyze these questions, and come to a conclusion. (Veillette, 2006)

Discussion:

Automated systems may be excellent in concept and in their implementation, but one must remember that automation also carried with it certain unforeseen dangers, problems and disadvantages. Automation is today used widely in every field, no matter how small or how large it may be. As far as the aviation industry is concerned, automation has penetrated the cockpit as well. According to the British Airline Pilot's Association, or the BALPA, automation may lead to problems for pilots as well. In their words, "Airline pilots increasingly lack the basics of the flying skills and may be unable to cope with an in-flight emergency such as sudden mechanical failure" as a direct result of the automation of the cockpit. (Veillette, 2006) This could also mean that pilots today lack the skills that are needed to fly the aircraft manually when and if the need were to arise. (Veillette, 2006)

Martin Alder, a member of the British Airline Pilot's Association stated, "The style of flying and training means that people will be less able or less likely to cope, which has obvious safety concerns." (Veillette, 2006) One must remember the fact that when the idea of cockpit automation was initially conceived of, several promises were made about the innate advantages that such a thing would ultimately have on the aviation industry. One such purported advantage was that the capacity of the national airspace system would be increased dramatically, because of the simple fact that the boxes would be able to make navigation much more precise than before. The manual workload of the pilots would be reduced as well, and this could be a good advantage for the pilots, as they would feel relief from the inherent stress of flying an aircraft. Other routine operations generally carried out by humans would be reduced, and this could mean better and greater efficiency and effectiveness in the long run. Not only would the management of the aircraft become a simple affair, but the possibility of human error would be minimized as well, and these could mean only good things for the aviation industry, felt experts. (Veillette, 2006)

Although the automation of the cockpit was eagerly expected and anticipated, it soon became apparent that several problems had been completely unanticipated. When the automated cockpits began everyday line operations, it was noticed that the automation was actually creating more errors than an average human being would. This was happening despite the grand promises that automation would be able to effectively reduce human errors to a large extent. In the environment of a terminal too, airport workers noticed that the workload in an automated cockpit seemed to be much higher than in a non-automated one, especially in the age old steam-gauge cockpits. Furthermore, automation had seemed to add to the problem of both pilots to go 'head down' while they were in busy terminal airspace, and this was a dangerous problem that could not be overlooked by any means. However, the most serious problem by far seemed to be that of the automation lulling pilots into a sort of complacency, in which they would take many things for granted, in the confidence that the automation would allow them to do so. Manual flying skills were also lost gradually over time, as more and more pilots started to depend on automation to take care of flying, and more and more flight crew managers started to express concern that they felt that pilots were losing their basic 'stick and rudder proficiency', a skill that would help them survive in case automation failed and they would be forced to manually resume direct control of the aircraft. (Veillette, 2006)

Take for example the accident that occurred due to engine failure on January 8, 1989, on board British Midland B. 737-400. When the outer panel of one blade on the left engine detached, the no 1 engine started to stall badly, and this resulted in aircraft shuddering and ingress of smoke to the flight deck. The crew, believing that it was the no 2 engine that had created these several problems, shut it down. This caused the surging of the no 1 engine to slow down considerably, leading the crew to believe that they had been able to deal with the emergency satisfactorily. They soon shut down the no 2 engine, and the no 1 engine operated properly for a while, although accompanied by shuddering. Soon however, there was an abrupt reduction of power, and a fire warning was issued. The efforts to restart the no 2 engine were not successful, and the aircraft crashed; thirty nine people on board the flight were killed instantly, while a further eight died in hospital later. Seventy four out of the seventy nine other occupants suffered injuries, and the cause of the accident, as established later, was that the operating crew shut down the no 2 engine after a fan blade had happened to fracture the no 1 engine. Would the accident have occurred if automation had not been carried out within the cockpit? ("Flight Deck automation issues," n. d.)

Experts do feel that cockpit automation may well lull pilots into a false sense of security, that nothing would go wrong, just because the cockpit is fully automated, and that the machines would take care of any adversity. This, feel experts, would make pilots more prone to making fatal errors in certain situations, which they would otherwise have been able to handle with elan, before automation. Recently, a study was carried out by Dr. Linda Skitka, an Associate Psychology Professor at the University of Illinois in Chicago. The study was aimed at studying cockpit tasks carried out with the aid of a computer, and without. It was found that the error rate among students was an amazing sixty five percent, with the students lulled into a false sense of complacency by the computer, which was giving them false prompts. The students followed the prompts of the computer as against following their own knowledge and awareness of the cockpit, and even though other instrument readings on the panel contraindicated the readings of the computer. Dr. Linda describes her experiment: she used a basic flight simulator and divided the eight students into two groups. While half the students were to fly with the help of an automated computer system, the other half were supposed to rely completely on instrument readings. Both the groups were told that the instruments were completely reliable, and those students who were flying with the automated system were told that although the automation was reliable, it was not a hundred percent so. ("Cockpit automation may bias decision making," 1999)

The basic idea behind the experimentation was to test out the fact of whether or not errors of omission and commission were carried out while flying. While errors of commission meant complying with an erroneous computer prompt, despite knowing that the instruments were providing contra indicatory information, errors of omission meant failing to respond on time to a correct computer prompt, which would be consistent with the information being displayed on the instrument panel. The experiment revealed that the six errors of omission were in fact constant between both automated and non-automated conditions within the cockpit, with the idea of testing whether automation would lead to a decrease in vigilance proving to be true. Dr. Skitka went on to state that automation was in fact preying on basic human weaknesses, like for example cognitive laziness, social loafing, a diffusion of responsibility, and so on. Cognitive laziness, according to the experts, is a condition in which people reveal a tendency to take short cuts for a number of things, including a short cut to flying on automation, as in this case. Social loafing refers to the tendency displayed by people, in which people tend to expend lesser effort in any given situation, when there is a group of individuals involved. ("Cockpit automation may bias decision making," 1999)

Most individuals, stated Dr. Skitka, tended to display a tendency to slack off when there was a situation that warranted a sharing of responsibility. Therefore, in this case, when the computer is a part of the group, the same tendency would apply. Diffusion of responsibility refers to the inclination of individuals to conform to the demands that figures of authority make on them. In this particular case, the computer was taken as an authority figure, or at the very least, a figure that was infinitely smarter than the user. The users of the computer aided automation within the cockpit were more likely to follow the instructions being given by the machine, even though it may be contra-indicatory or contradictory. As a direct result, introducing a computer to make automated decisions for pilots may well lead to the creation of an entirely new and different set of errors, as in the Doctor's words, "Changing the context changes the opportunity for error." ("Cockpit automation may bias decision making," 1999)

According to Donald a Norman, of the University of California, San Diego, automation in itself is not the problem; rather, it is the inappropriate use of automation that is causing all the problems in the aviation industry. In his opinion, inappropriate application of automation could well lead to accidents of the kind that the aviation industry tries to blame on 'over-automation' and its associated pitfalls. Therefore, he feels, automation must be brought in to the industry today, although within a more appropriate framework, or at least by making a concerted attempt to remove some forms of already existing automation. Today's automations have 'an intermediate level of intelligence', which tends to aggravate the existing problems, and to maximize difficulties and problems. The design of the automation can also be extremely important, feels Donald Norman; the design must be able to encompass the entire system, that is, the equipment, the crew, the social structure within the cockpit, the training and learning activities, the cooperative activities carried out within the framework, and also the overall goals of the task. (Norman, 1990)

More often than not, automation tends to be implemented with absolutely no changes within the overall design first, and this could mean that the automation would be a failure rather than a success, which could have been avoided if efforts had been made to improve the design at the outset, before bringing in automation. Take for example the tasks of the crew on a commercial airline, where almost all the flight activity can be considered to be normal and routine. It is common knowledge that large modern aircraft are easy to fly, and the aircraft would generally be maneuverable, stable and responsive at any time, with automation automatically monitoring all in-flight equipment and operations thereby easing the workload of the crew. This is evidenced in the fact that today's aircrafts require only two people to fly them, although in reality only one person would be needed, as compared to the aircraft of yore in which three people were expected to be present to fly the aircraft. (Norman, 1990)

Today, there are lesser numbers of reported accidents, and it is generally assumed today that the decrease can be totally attributed to the fact that automation has pervaded the industry completely of late, and this means that it can be safer today to fly an aircraft. This may or may not be true, feels Donald Norman. For one, the crew flying a large plane may not in actuality be aware of all that is happening on the flight; they may be physically isolated from the passenger section of the flight, and may not know what is actually happening there. The crew is also more often than not isolated from the physical structures of the aircraft, and to add to this, the mental isolation that the crew faces as a result of the automation within the cockpit can in fact lead to greater danger than ever before. This is because of the fact that the automated controls within the cockpit that is responsible for monitoring and controlling the aircraft leaves no room or trace of its various operations to the crew managing them, and this would isolate and separate them from the moment to moment operations of the aircraft and the control within the cockpit. While it may be true that this form of isolation or segregation would be able to effectively reduce the workload of the crew, and also the reliance on human variability and failures, it may also contribute directly to the magnitude of the problems if any within the aircraft, when the crew is faced with immediately diagnosing the escalating situation, and in designing an appropriately best course of action with which to face the situation. (Norman, 1990)

Experts state that physical isolation as such would be suitable for the crew, if they kept themselves up-to-date on the critical states and stages of the devices that they are controlling and manipulating, but the problem with physical isolation and separation could be that it would automatically lead to a form of mental isolation as well, and when this happens, it could be very dangerous indeed. Zuboff gives an example of the control room of a modern day paper mill. Whereas in previous times the crew would remain on the floor, and therefore aware at any moment of what was happening to the equipment, now the crew may be poised on the floor above the mill and other equipment, sitting isolated within a sound proof glass covered cabin, with no real physical evidence of what is happening down the stairs away from where they are sitting. This would make them vulnerable to mistakes and errors, especially since they are removed from the meters and other displays of the mill while work is in progress. In the same way, automatic equipment on an aircraft would effectively isolate the crew from the actual functioning of the machines within the cockpit, and this could be dangerous indeed in any given situation. (Norman, 1990)

Take for example the case of Flight China Airlines 747, which rolled and went into a vertical dive from about 31,500 feet, severely damaging the aircraft and leaving no room at all for recovery. What happened was that the aircraft suffered a gradual loss of power from its right engine, and when normally this would have caused the aircraft to yaw to the right side, the autopilot compensated for the yaw, until such time that it reached the final stage of its ability to compensate for the yawing, and the plane could no longer remain stable. It was at this time that the crew was able to determine the actual problem, but by this time, of course, it was much too late to take a proper decision about the action to take to make up for the loss of power from the right engine. (Norman, 1990)

The aircraft went into a roll and nosedived from several thousand feet. The questions is this: could this accident have been averted if the human crew manning the aircraft had been more vigilant and careful, rather than rely completely on the automations that had been implemented within the cockpit / could this accident have been stopped from happening, if a human being had been manning the aircraft and manipulating the controls, rather than the autopilot than took over when it could and stopped when it could go no further? Take the other example of a vigilant officer managing to detect one problem, while at the same time failing to detect another on his aircraft. The second officer reported that although he was feeding fuel to all the three engines of the aircraft from the number 2 fuel tank, the number 3 fuel tank was showing a decline in its fuel. Soon enough, it was noticed that the wheel was cocked to the right. The pilot was instructed to turn off the autopilot, and when this was done, it became apparent that the aircraft was displaying a 'roll' tendency, thereby signifying the reality that they were now faced with an out of balance situation. (Norman, 1990) large amount of fuel was being lost, and the aircraft was in imminent danger of crashing. In this example, it is obvious that it was because the second officer was able to provide the very valuable information that there was something horribly wrong with the fuel balance and this despite the fact that the autopilot had managed to quickly and efficiently take over to compensate for the resulting weight imbalance on the aircraft. The question here is this: if the autopilot had been able to signal to the crew that something was wrong with the fuel balance, because it was compensating more than it normally did, would this have been able to alert the crew to a potential problem much quicker? Although it may be true that automation did manage to save the day, and perhaps if not for the autopilot taking over and managing the fuel balance, the aircraft would have crashed, there have been several arguments about automation, and in general opinion, it would be a good idea to use less automation and rely more on good old human man power. (Norman, 1990)