Twenty-five Years of Flight Simulation
- or -
It's a Long Time Since I Lost a Buddy in a Training Accident.
by John Rolfe
Twenty five years in no way encompasses the history of flight simulation, so why chose this time to look back? Two events prompted this article. The first was that in 1970 the Royal Aeronautical Society held its first conference devoted to flight simulation. The success of the conference generated interest and support for the formation of the Society's Flight Simulation Group. Since then the Group has maintained a highly successful international conference programme. Moreover, the Group has taken a number of productive initiatives to enhance understanding and encourage progress in the design and use of flight simulation. This article draws on the activities of the Group to consider the changes that have taken place and the progress that has been achieved.
The second reason for the title was that in 1971 Flight International published an article from me entitled 'The Future of Simulation'. In it I concluded that simulation, at a range of levels, had the potential to play a much greater role in aircrew training. To achieve this objective would require addressing the issue of motivating both users and legislators to accept simulation as a means acquiring and maintaining competency. Twenty-five years later I will attempt to assess how accurate I was.
The role of flight simulation has changed significantly over the last twenty-five years. In 1970 flight simulation was very much a subordinate to training undertaken in the aircraft. However, this was the time when the NASA Apollo Moon Landing Program was in progress and the key role of simulation as the means of providing total mission training was being widely publicised. The Apollo simulations were seen to be successful because, in the words of Murray, and Blycox in their book ' APOLLO The Race to the Moon' (Simon & Schuster, 1989), they met four requirements
"First, they were a way of keeping in shape, the equivalent of fivefinger exercises for a pianist. Second, they gave everyone a chance to see whether the mission rules and techniques that worked on paper also worked when put to a test. Third, they gave the controller experience of how to deal with the particular set of emergencies portrayed in a particular simulation. And fourth, they made the controllers familiar with fear, putting them through such hair-raising, gut-wrenching adventures that nothing they encountered in an actual flight could seem more terrifying or hopeless."
While the civilian and military operators could look to the Apollo Programme, and envisage a wider role for flight simulation, they had different objectives for the directions in which they saw their policy advancing.
The civilian users reasons for wanting a greater training role for flight simulation were clear. Training in the air took costly resources away from their revenue earning role of carrying payloads. Moreover, it could be inefficient and dangerous. At that time the bizarre statistic was that, for some aircraft types, more serious in-flight accidents were occurring while practising certain emergencies, than resulted as a consequence of the emergency ever actually taking place. It was in this context that an airline's representative speaking at the Royal Aeronautical Society in 1970, put forward the following policy statement;
"The maximum use of Flight Simulators is advocated in order to reduce the amount of in-flight training required for initial and recurrent training, with the objective of totally removing the need for in-flight training exercises, apart from route training, to be carried out in the aircraft. All potentially hazardous training manoeuvres should be transferred from aircraft to flight simulators, as and when these can be satisfactorily accomplished in flight simulators."
Embedded in these objectives was the idea of Zero Flight Time Training (ZFTT), i.e., that all conversion training, either from one aircraft type to another or from one flight crew position to another, would be carried out either in flight simulators or during routine revenue earning flight operations. A quarter-century later ZFTT is now an accredited and established feature of commercial aircrew training and certification.
The adoption of ZFTT has been in cautious steps rather than a revolutionary change to the pattern of commercial aircrew training. Some major airlines have only recently achieved the transition to ZFTT. Others, however, have been able to cite some remarkable statistics for the impact of simulation on their aircrew training programmes. For example, from January 1981 to December 1990, American Airlines successfully trained 13,853 pilots in total simulation programmes.
Civilian flight training has been changed by the introduction and acceptance of flight simulation. No doubt, at the present time, this change is will be identified with the added sophistication that flight simulation has contributed to airlines' training capabilities, e.g., its role in CRM and LOFT training. But, equally significant must be the recognition that airline training is no longer a hazardous activity. This fact is eloquently reflected in the comment by a senior training captain with a major US airline at the RAeS in 1996. He said, "For me the impact of flight simulation over the last twenty-five years has been that it is a long time since I lost a buddy in a training accident!"
In 1970 the military users of flight simulation were in a more ambiguous position. Unlike their civilian counterparts the great proportion of military flying in peace time is to undertake some aspect of training, i.e., developing or exercising operational capabilities. Consequently while they embraced a policy for the greater use of simulation they were also adamant that for military aviation there was no substitute for actual "hands-on" training in the air. A key reason for this apparently conflicting position was that cost pressures on the military were forcing the pace for the greater use of simulation in training. Consequently, simulation was seen as being imposed upon the military users rather than being the preferred course of action.
As commercial aviation had ZFTT as its potential encumbrance the military were creating theirs in the guise of Full Mission Simulation (FMS). With this aim in mind, a paper from the MOD presented to the RAeS in 1970 confidently stated:
"The RAF is introducing a new generation of flight simulators known as 'Full Mission Simulators'. These simulators employ the latest electronic techniques, and will be more advanced than any military simulator in the western world."
As the subsequent decade was to show, while FMS continued to be widely espoused, its implementation was to fall short of many users' expectations. This was reflected in the contributions from the RAF to subsequent RAeS conferences.
"Although the term 'Full Mission Simulator' has been bandied About for years, the concept has not been fully realised yet. Doubts about the way the RAF was approaching simulation were first voiced in the early 1970s. There was at that time no official simulator policy. As a consequence the RAF was making decisions based entirely on subjective assessments and manufacturers claims. Not surprisingly a number of simulators failed to live up to expectations and we compounded this problem by directing our aircrew to do more simulator time than was justified by the training value available, thus creating understandable resistance."
"It is a sad fact that most of our fast-jet simulators have inadequate handling fidelity. Indeed there are many senior Air Force officers, mainly with a fast-jet background, who, having been weaned on so-called 'Mission Simulators' have never forgotten the claims made for these machines nor their actual capabilities. So we have an attitude barrier to break down and this is proving difficult."
In fairness to the military users their training requirements have been exceedingly demanding upon themselves as well as flight simulation. In adapting to both the changing state of military aviation and training technology military aviation has shown itself to be flexible and willing to:
(i) Accept the idea that, for some training tasks, part-task trainers can be as effective as full simulation . Part-task trainers range are used in basic flying training through to the operational level.
(ii) Introduce simulation at all levels of training from the selection candidates for pilot training through to mission rehearsal and evaluation at the operational level.
(iii) Change its position and to adopt simulation applications where previously it was considered either not feasible or viable. Examples are air combat and air-to-air refuelling simulations.
(iv) Take high technology approaches where necessary. Examples are the RAF's use of Area of Interest visual displays, and the United States Department of Defense's development of the SIMNET networked tactical simulation system.
The enhanced role of flight simulation has been supported by technological change. In 1970 the future for simulation looked towards the development of hybrid analog/digital computers. Then few would have anticipated the dramatic changes in computing techniques and capacity that have taken place. The digital computer brought change to the hardware and the architecture of simulation. Large computing systems have given way to distributed processing systems. Increased computing capacity has allowed greater sophistication of the simulations as well as the host simulators.
The change in simulation capability as a result of increased computing power is exemplified by the capacity to simulate the external visual world. Twenty-five years ago training simulators were fitted with visual systems in which a small television camera moved over a scale model containing significant, but restricted, features of the world, e.g., airports and potential military targets. The most likely means of advancing the capability of visual systems were seen as combinations of laser scanned phototransparencies, motion picture systems and limited use of computer generated dusk/night images of airfields for take-off and landing.
Today the digitally generated visual scene has a virtual monopoly of the means of creating the representation of the external world. This has been achieved by means of advances in computing power, computer graphics technology and the availability of geographic information in digital form.
The area in which technological change continues to be sought is the means of presenting the image. Until now the means of doing so has been the cathode ray tube which still suffers from limited operating lifetime and low light output. Alternative solutions are being explored for example, liquid crystal displays and digital micromirrors.
A visual cue that remains absent from current simulators, is changes in ambient light and shadow in the cockpit as the aircraft's position alters in relation to the sun, a very powerful indication of change of orientation and attitude.
While the improvements to the representation of the external visual scene has been generally endorsed the value of providing motion cues has continued to be the source of much debate. Critics of motion argue that a simulator is constrained by being fixed to the ground and consequently no motion system will be able to reproduce the full dynamic response of the aircraft. Moreover, motion systems greatly increase the cost of the simulators, their accommodation and maintenance. They further support their case by presenting evidence that wide field of view visual systems provide sufficient motion cues. Those in favour of motion argue that, because the flight environment is inherently dynamic, a fixed base simulation is an inadequate representation of flight. Moreover, experimental studies have shown that aircrew behaviour in a simulator with motion is much closer to that occurring in actual flight. It is, therefore, better to have some motion than none at all.
One very important contribution to this debate is the distinction between feedback and disturbance motion cues. Feedback motion is the representation of dynamic changes to the aircraft resulting from actions initiated by the pilot in control. An example of feedback motion would be the onset of rolling motion as the pilot banks the aircraft in order to turn. Disturbance motion represents the dynamic changes to the state of the aircraft resulting from the effects of external factors, e.g., turbulence which can cause the violent transient or prolonged displacement of the aircraft from its intended attitude and course.
Disturbance motion cues are those which need to be represented and their absence from a simulation may result in skills being acquired that are inadequate in the real world when actual disturbance motion is encountered. The difficulty is that the ability to represent disturbance motion accurately can require very complex motion platforms with capabilities for large and sustained amplitude/acceleration responses. These qualities may be present in one or two of the most sophisticated flight simulation research facilities but not in off the shelf motion platforms fitted to training simulators where cost, space and complexity would inhibit their provision.
In the fields of simulation technology just as the flight simulator users have had their 'quantum jump' concepts such as Zero Flight Time Training and Full Mission Simulation so too has the industry. It has generated its own collection of 'significant innovations' that have come and gone, often taking their creators with them. The buzz words of the moment are Virtual Reality. Without doubt VR does contain some new ideas and technology, but it has also been guilty of ignoring the achievements of simulation or, on occasions, been willing to take existing technology and invest it with the VR label and a "value added" price ticket.
Another area in which changes have taken place is the role of the simulator instructor, instructor work-stations and performance recording and display systems. Twenty-five years have seen great changes to the facilities provided for simulator instructors along with their functions. In 1970 the simulator instructor was expected to play a major part in generating 'live' simulation training scenarios while carrying out the functions of instructing. Very often this meant that the instructors' station was placed closer to the computer than the crew under training. Advances in computing and training technology have relieved the instructor of much of the machine-minding. The 'on board' instructor station is now the norm, while the capacity to draw upon a library of preplanned automated scenarios allows the instructor to devote more attention to supporting training.
In 1970 objective performance measurement was not in wide use. In fact there was debate whether crews would accept their performance being measured in depth by computer-based methods. Speaking at the RAeS Derek Ruben, from what was then the Air Registration Board, argued that it should be possible to measure at least the same number of parameters in the flight simulator as were then being recorded by aircraft flight data recorders.
Data recording is now an accepted part of a flight simulation facility with, in addition, the ability to analyse and present the information obtained in the most appropriate and user beneficial manner. Military simulation has made use of this capacity with the adoption of debriefing facilities allowing crews to replay and assess their performance outside the simulator.
Along with data recording and presentation there is the capability to interchange data between simulators (i.e. networking) and between the simulator and the aircraft. The latter facility permits the use airborne data as a source from which to generate specific flight situations for both training and evaluation purposes, e.g., aircraft accident investigation.
The desire of users to increase the role of flight simulation, and the growth of technology to improve the capabilities of the simulators, have had to be harmonised with the attitudes of the legislators. Back in 1970 Captain Johnson from BEA reminded the RAeS that:
"A statement of policy is one thing but putting it into practice is much more difficult, especially when Government authorities have to be convinced that the flight simulator can be used to do the training as effectively, and in some areas, more effectively than the aircraft."
There was also the reminder from the Air Registration Board that, because of the nature of commercial aviation, any standards that were drawn up needed to be internationally approved and followed.
It is timely to recognise the significant role the RAeS Flight Simulation Group has played in establishing international standards for flight simulation. In 1989 it set up an international working group with representation from industry, airline users, and national regulatory bodies. After meetings in the USA and Europe, a document "International Standards for the Qualification of Airplane Flight Simulators" was approved at the RAeS in 1992 that defines the performance and documentation requirements for the evaluation of flight simulators used for training and checking of cockpit crew-members. The value of this initiative may be measured by the fact that the standards have been accepted by the International Civil Aviation Organisation (ICAO), and the European Joint Aviation Authorities (JAA).
Following on from this initiative the Group has considered the methods by which the assessment of flight simulators is carried out. This work has recently been published by the RAeS as "The Airplane Flight Simulator Evaluation Handbook."
Throughout this time, when enhancements to the role and capabilities of flight simulation have been taking place, two performance related demands have been made on users. These are to show evidence of cost and training effectiveness.
Back in 1970 Captain Jack Nicholl then with BOAC, stated the position of the commercial aviation regarding the use of flight simulators when he said
"If simulators did not save the airlines money they would not exist, and, however valid the other reasons are, we would have no difficulty in convincing ourselves that training in the air was better in every respect!"
In 1989 Brian Hampson from the simulator industry put its position as follows:
"Manufacturers are not philanthropic organisations whose role is to provide the ultimate training device at little or no cost to the purchaser. Unlike all the other players in the acquisition process, the sole reason for the manufacturer being in the game is to make money. ..He is not directly, nor primarily, concerned in reducing the costs of training. These are all the primary concerns of the operators and the Regulators."
These are tough criteria and their impact has been felt by both sides over the intervening years. For the flight simulator users changes in aircraft operating costs, reductions in crew numbers and increased aircraft sophistication have lead them to seek more return on their investment in flight simulation.
The much quoted rule of thumb used to be that the aircraft to flight simulator operating costs were of the ratio of 101. This no longer remains true as the operating costs of the aircraft reduce and increased simulator sophistication have to paid for. For the new generation transport aircraft such as the Boeing 757/767 the ratio has been put at 41. Moreover, Edward Boothe, who was until recently the FAA's expert on flight simulation, speaking at the RAeS cautioned that with commercial airplanes that cost less the ratio changes significantly. For a Canadair RJ or an ATR 72 the ratio approaches 11 while for a Jetstream 41 it is about 12 and for a Beech 1900D it is about 13. These adverse cost ratios have a significant impact upon the willingness of regional carriers to transfer their training away from the aircraft. They also restrict the ability of the legislative authorities to make the use of simulations a mandatory regulation at this level of operations.
Consequently, the users have had to take a more critical approach to assessing the benefits to be derived from the purchase and operation of flight simulators. The result has been that the customers have become more exacting in their specifications. As one airline's speaker recently commented:
"Very often when I get demonstrations of new features or new improvements in simulators I ask what this improvement is for, what is the training value? Most often, if I get an answer, it is that the technology has now made it possible. That is not good enough!"
The resulting impact on the flight simulator industry has been a reduction in the number of major manufacturers, along with a steady increase in new entrant suppliers offering alternative training devices, e.g. the growth of part-task trainers.
The debate over the need to measure the effectiveness of simulator training has continued unabated over the last twenty-five years. There are those who argue that, despite the increased ability to produce better simulators, there continues to be a requirement to provide objective evidence that flight simulation does actually produce the training results expected of it. Opposing this is the school of thought that says flight simulation has already proven that it can deliver transferable training and no further proof is necessary. Unfortunately, for the latter argument, in the high investment areas of new forms of simulations, (e.g., networked simulations and Virtual Reality) the issue is not whether transfer will occur. Rather, it is to know how much transfer and for what training objective, e.g., skill acquisition, maintenance or assessment.
One of the prevailing philosophies of the past 25 years has been that if there is a problem then a more complex simulator is the answer. This may be true when the problem relates to device realism, e.g., the representation of the visual world. But it is not necessarily so when the issues relate to device function, e.g., the effectiveness of simulation in training. The following words of caution from Prophet, Caro and Hall, written in 1971, are still valid and worth reflecting upon:
"If the aircraft is a poor learning environment, and for many flight related skills it is one of the poorest imaginable - then a ground based duplicate of that environment will not necessarily be a better one in which to learn."
It has to be recognised that the requirement for total simulation is driven very firmly by the prerequisite that regulatory performance assessment should be carried out in as representative a flight environment as possible. No matter how valid this objectives is it must not be allowed to obscure the fact that the acquisition and maintenance of aircrew competencies can be undertaken in devices that do not attempt to replicate the flight environment in every detail.
With the growing range of flight training devices that are on offer the danger is that, in situations where the performance of products can have an impact on the well-being and safety of others, it may regrettably be the case that assumptions and claims for training effectiveness will be tested by legal actions in the courts, rather than in the pages of learned journals. If this turns out to be so then the rigour of the evidence will certainly be put to the test under these circumstances.
So much for the past, what of the future? If lessons from the past are sought, the following suggestions are offered with respect to our policy and aspirations for flight simulation.
* It pays to be optimistic - if the user's training objectives can be stated, they are likely to be met.
* It is hazardous to predict the rate and direction of technological development.
* Be wary of conceptual 'milestones' that carry innovative labels as they frequently become either millstones that handicap their creators or media buzzwords which lack substance
* If a problem exists relating to the use of simulation the introduction of more sophisticated simulator technology is not likely to solve it.
The principal lesson from the last twenty-five years has been that progress may be inevitable but effectiveness is not. The progress of flight simulation has demonstrated very clearly that its successful use depends upon the synchrony of three factors namely:
1. Motivation - the users must have the will to want to adopt simulation.
2. Innovation - the industry must be able to provide the quality of simulation the users require.
3. Opportunity - there must be the regulatory flexibility to encourage and permit the wider application to simulation.
These three factors do not automatically act in concert with each other. Very often it requires special initiatives and opportunities for those involved to be able to come together to explain their positions and, hopefully, agree a policy for the future. The RAeS Flight Simulation Group is to be congratulated for undertaking this function so effectively over the last quarter-century and wished success with its continuing activities.
About the author.
John Rolfe is a visiting professor at the College of Aeronautics, Cranfield University. At Cranfield he is involved in courses on flight simulation and developing crisis management simulations for the training of aircraft accident investigators and airline operators. He was the first Chairman of the RAeS Flight Simulation Group and remains a committee member.