History

Digital Simulators

One of the restrictions in these early days was that aircraft manufacturers did not have much analytical information on the performance of their airframes and engines; the simulator manufacturers were therefore required to use ad hoc methods to achieve the desired aeroplane characteristics. This changed however, with the arrival of the large subsonic jet transport era when the aircraft manufacturers began to produce much more complete data and to perform more extensive flight development programmes. Together with requirements for driving the motion and visual systems then being introduced and pressure from the operators to improve accuracy and thereby, they hoped, better transfer of training, significant increases in the amount of analogue computer hardware became necessary to satisfy them. At this point, the law of diminishing return began to operate, the cumulative errors caused by all the additional hardware exceeded the improved accuracy which should have resulted from the more extensive aircraft data, which demanded the extra hardware.

In addition, reliability began to fall in spite of improved hardware and design technology, or at best was only maintained by the efforts of maintenance teams. At that time, the required utilisation was around 8-10 hours per day for five days per week. This was soon extended to six days per week, even then, the requirement of today, for a training utilisation of virtually 24 hours per day for seven days a week could be foreseen. It thus became obvious that the demands for increased fidelity of simulation and reliability could no longer easily be met with analogue machines even with the use of the new solid state elements which had appeared. Around this time the second generation of digital computers, started to materialise, and were able to satisfy the speed and cost requirements of flight simulation. As a consequence, there was an almost total swing to digital simulation for all but the simplest trainers.

It was realised from the earliest days of programmable electronic digital computers that a potential application would be in real-time digital simulation. The advantages of digital computers, improved flexibility, repeatability and standardisation, were approached by the U.S. Navy who initiated a research program at the University of Pennsylvania in 1950. The general purpose computers of the time could not be used directly for real-time flight simulation, due to their poor arithmetic and input-output capabilities. A special machine therefore, was designed at the University for their simulator, which was named UDOFT (Universal Digital Operational Flight Trainer). This computer was manufactured by the Sylvania Corporation and completed in 1960. The UDOFT project had demonstrated the feasibility of digital simulation and was mainly concerned with the solution of the aircraft dynamic equations. In the early 1960's Link developed a special purpose digital computer, the Link Mark I, designed for real- time simulation. This machine had three parallel processors for arithmetic, function generation, and radio station selection. In the late 1960's general purpose digital computers designed for process control applications were found to be suitable for simulation, with its large input - output requirement, and the use of special purpose machines declined. Today special purpose digital computers are only used in applications demanding very high speed processing, such as computer generated imagery.

Nearly all of the simulators produced up to the mid 1950's had no fuselage motion systems. This was justified by the statement that modern pilots did not fly "by the seat of their pants", but the fact remained that fixed-base simulators did not feel like aeroplanes to fly. It was found that a handling improvement could be made by empirical adjustment of the control loading and aircraft dynamics simulations which, in part compensated for the lack of motion. Proposals were made by the manufacturers for motion systems, but it was not until the late 1950's that the airlines decided to purchase them.

In 1958, Redifon received a contract from BOAC for the production of a pitch motion system as part of a Comet IV simulator. More complex motion systems were designed capable of producing motions in two and three degrees of freedom, and with the introduction of wide-bodied transport aircraft, such as the 747, a lateral acceleration was required which led to the introduction of four and six degrees of freedom systems. Six degree of freedom motion systems are now the most common. The perception of motion and its effect on training is one of the less understood aspects of simulation and research is still active in this area.

Systems for producing the extra-cockpit visual scene have been proposed and constructed for almost as long as flight simulators themselves. However, realistic and flexible visual attachments are a fairly recent development. Due to the large number of visual systems which have been invented, only some of the more successful ones can be mentioned here.

The point-light source projection, or shadow graph, method enjoyed popularity in the 1950's, especially for helicopter simulators. A series of simulators using this method of visual display were produced by Giravions Dorand in France including an ab initio hovering trainer produced by Shorts of Belfast in 1955. Simulators on this pattern were also built in the United States, but the shortcomings of the shadow graph system seems to have limited the success of the concept. The first visual systems achieving widespread use on civil aviation simulators were based on the scale model and television camera method, although methods based on film and anamorphic optical systems have also met with success for more restricted applications. Serious development of closed-circuit television visual systems began in the mid 1950's with monochrome systems being produced by Curtiss-Wright, Link (then the Link division of General Precision) and General Precision Systems (formerly Air Trainers and Air Trainers Link Limited). The first colour system was produced by Redifon in 1962. Television based visual systems have under gone a steady development since then, with a large part of the effort being devoted to improved methods of image display.

The first computer image generation systems for simulation were produced by the General Electric Company (USA) for the space programme. Early versions of these systems produced a patterned "ground plane" image, while later systems were able to generate images of three-dimensional objects. Progress in this technology has been rapid and closely linked to developments in digital computer hardware technology. Current systems available from major simulator manufacturers are able to produce full colour images with scene contents of several thousand polygons and point-light sources. A parallel development has taken place in night-only computer image generation systems; these use the calligraphic or stroke-writing, rather than the raster scan method of display, which enables a superior reproduction of light points. The first of these systems was produced by the McDonnell-Douglas Electronics Corporation in 1971 and called Vital II.

Current systems, can produce images of night, dusk, and daylight, using resolutions of up to 3 million pixels, and 20,000 surfaces, as well as 4000 calligraphic light points which are plotted during the flyback time. They can employ texture maps, photographic images of real ground detail, pixel level fog effects, texturing, height above terrain information, and up to 256 moving models. All of this is per channel, of which there may be 3 or 5, on a typical simulator, and all manipulated in real time, at frame update rates from 30 to 70Hz.

A Computer generated view of Frankfurt Airport at dusk.

Much effort has been devoted to improving the instructional facilities in the simulator. The use of high resolution touch screens for instructor control, and substantial increases in the number of malfunctions and radio stations which can be offered, there are also facilities for exercise recording and playback, pilot performance recording and evaluation, separate pilot and flight engineer training in the same exercise and automated training. We have now reached a point in commercial flying training, where all conversion and recurrent training can be conducted in a simulator, so that a pilot of one aircraft type, can be cross trained to another, without ever actually having flown the real target aircraft, until he or she is on board, carrying fare paying passengers.