Ivan Sutherland's Sketchpad in 1963 is often referred to as the first milestone in computer graphics. It was a line drawing, two dimensional system that flickered from the slow refresh capacity of a random scan display. Yet in this first seminal effort in computer graphics, we also could have observed a har-binger of the field of graphical simulation and design. The demonstrations in Sketchpad were of a design nature signaling to future researchers the dividends that computer graphics and simulation could bring. But only today are we entering the greatly heralded era of using simulation extensively to aid the design process. Hence this conference is very timely. The following papers show in technical detail some of the ways in which simulation will be used to extend the creative design process.
Some of the well-known methods of computer graphics, especially those useful in simulation, are reviewed qualitatively, to serve as an introduction for readers in less related fields, to the succeeding papers. Operations necessary to display points, lines, polygons, and polyhedra, both stick-figure and solid-face, including hidden line and surface removal, are discussed.
A geometrical graphics system will be described and illustrated with slides. The system is implemented at the Pacific Missile Test Center, Point Mugu, California, utilizing an IBM 7094 computer and a DatagraphiX 4460 COM recorder in a batch processing environment. The system features small core requirements, dynamic storage allocation, ANSI FORTRAN subprograms, device independence, and internal homogeneous coordinate techniques. User capabilities include n-dimensional constructions, prism and pyramid generation,cross sec-tioning and truncation of solids, automatic view point selection along an axis and automatic scaling, implied edges, hidden line deletion, and transformations such as dilation, translation, rotation, reflection, and projection. Graphical output presented will include line drawings of esthetic and commercial art experiments, crystallographic solids, and other geometric illustration.
This paper discusses the traditional design process as it often applies to geometric design, and then it examines a new organization of the design process in which computer simulation plays a central role. Some future developments in computer aided geometric design are anticipated.
Computer animation can be implemented using parametric polygonal models. A parametric model consists of a set of polygons, a set of functions and a set of parameters. The position of each polygon vertex is determined as a function of one or more of the para-meters. With parametric models, animation is reduced to specifying the parameter values for each frame of a sequence. Two models are used to illustrate this concept.
The development of real time scene generation for simulating a real world environment has come a long way since its modest inception over three decades ago. Nonprogrammed real time scene generation systems date back as early as 1939 when a simple cloud simulator was attached to a Link AN-T-18 basic flight trainer. The cloud simulator was comprised of a point light source which projected cloud images from a film transparency onto a screen. Since this simple beginning, point light source techniques have been refined, optical image generation systems have been devised, semiprogrammed motion picture systems have been produced, and complex television systems with camera-model, film, and computer generated imagery have been developed. As a result, current scene generation technology provides the capability to simulate in real time a real-world visual scene for most commercial and military vehicle operator tasks.
Simulated images have been used for training for at least thirty years. This has included simulation of visual scenes and of displays from various sensor systems such as radar, sonar, infrared, and television cameras. Techniques used have included scale models, acoustic tanks, and photography. Increasingly through the years, analog and digital computation equipment has been used along with these techniques for control and modification of results. In the early sixties, the increasing power of digital computation and data processing circuitry led to consideration of systems in which the "environment" hills, targets, buildings, etc. exists only as stored numbers, and images are generated in real time by appropriate computations performed on these numbers. Such approaches promised many advantages as compared with traditional techniques. Continuous effort has been applied since that time to develop algorithms, overcome disadvantages, improve performance, and reduce cost. The results: extremely realistic systems are now in operation. Recent radar simulation systems are based on computed displays. Visual scene simulation is proving itself on several operating systems. These have been combined in correlated multi-mode displays. Development is under way to apply this technology to simulation of displays from other sensors.
The LinkNVS tm(Link Night Visual System) is a compact, low-cost flight simulator attachment which generates visual scenes. The visual images are computed at 30 frames per second by digital computational implementation. The visual scenes are displayed by means of a color CRT operating in the stroke writing (calligraphic) mode. The computed display changes dynamically in concert with the real-time progress of the simulated aircraft through its flight plan. The NVS can be configured to provide up to eight independent view channels by means of subsystem modules. Each of the eight windows will display up to 2,000 light points computed for the dynamic airport display scene.
Most visual display systems for piloted simulators provide out-the-window monochromatic or chromatic scenes by means of televised views of airport models. A more recent and promising technique is the use of digital computers to generate the outside display scene. This technique has suffered, however, from limitations in perceived realism. This paper describes a calligraphic, chromatic system developed for improving the perceived realism of out-the-window display scenes, presents the results of a brief study to demonstrate such improvement, and discusses current developmental work and future research. The method selected for providing the natural color cues normally available to the pilot during a real night landing approach required the construction of a new chromatic projector for use with a calligraphic display. The chromatic projector permits (i) drawing 2000 vectors in as many as 500 colors, all above critical flicker frequencies, and (ii) using high scene resolution and brightness at an acceptable level to the pilot, within the maximum system capabilities of 1000 lines and 1000 fL (3426.26 cd/m2). In support of this new digital, calligraphic, chromatic projector system, a brief experimental investigation was performed in a fixed-base simulator to determine the effect, on system performance and pilot opinion, of a chromatic landing display as compared to that of a, monochrome display. Both performance measures and pilot opinion support the hypothesis that using a chromatic landing display improves system performance.
A feasibility model simulation system utilizing holography is described for use in training pilots to land on the deck of an aircraft carrier. The unique aspect of this system is that the visual display information is created solely by the hologram. The visual presentation is controlled by inputs from a simulated cockpit with a complete six degree of freedom motion system and a computer input for aerodynamics.
Most training devices require a visual system to present an out-the-window scene to the trainee. This session is devoted to two important subsystems of a visual system. These are: 1) Image pickup of model terrain and airfield by a servoed optical probe and camera assembly. 2) Display of visual image to the trainee.
Training simulators include displays of several types, using a variety of techniques. These are categorized as traditional, enhanced and visual scenes. Examples of these displays are described. The computation required in the simulator is contrasted with the process employed in the actual aircraft.
An optical scanning probe is essentially a TV camera lens that provides three angular motions of the line of sight through motions of internal components. Although probes usually have an extremely small entrance pupil, on the order of one millimeter or less, the unique requirements of their use with scale models in simulators causes a significant optical design problem. The designer must provide high image quality in a system operating over a broad spectral range, with a large field of view, having unusual re-quirements for extreme depth of field. These problems and some of the solutions that have been achieved are presented.
In flight simulation visual systems in which an image is picked up by a CCTV camera probe from a model terrain, there is a well known problem in focusing on the model surface across the whole imaged area, particularly as the probe entrance pupil approaches close to the model. The problem will increase as we try both to improve the resolution of the CCTV camera and probe and to reduce the scale of the terrain model, so that conventional techniques of moving and tilting lenses within the probe optics will not be sufficiently powerful for some desirable probe systems. There is a closely-related problem in correction of optical aberrations of probe systems for a wide range of object conjugate distances. The problems are assessed in relation both to normal TV systems and to laser TV systems. A novel form of tilt-lens system is described.*
The ability of television to display a completely flexible format together with the new ability to use digital computers to directly synthesize television video without use of cameras has created new applications for television projectors. These include: Command and control interactive displays both strategic and tactical. Simulation and training displays which duplicate the real world in color and perspective. Dynamic scoreboard, close-ups, and instant replay displays at concert and sports arenas. In addition, the visual impact of large screen commercial and closed circuit color TV together with rapidly declining projector cost is creating a vast home market which may rival the automobile industry.
This new color television light-valve projector consists of a deformable fluid control layer and electron gun in a sealed vacuum tube, an external light source and projection lens. It is being used for large screen color television applications such as simulation, entertainment, and control centers. Continued development of this type of light-valve projector has resulted in improvements in resolu-tion, sealed light-valve life, higher TV line standards, and higher brightness. Resolution gains in both color and black and white are due to refined optics and in applying new concepts for the "grating" modulators. The life of the sealed light valve has been extended to over 6500 hours by improvements in the cathode, ion pump and fluid. Operation at the 1029 TV line standards has been achieved for the monochrome light valve and prospects look good for pushing the single gun color projector to the same limit. Recent experiments have shown the feasibility of creating a "wide screen" format for either color or monochrome by the addition of an anamorphic lens and supplying the projector with a "squeezed" format picture signal. Applications of these projectors to simulation and other fields will be discussed.
The fundamental differences between Field Sequential and Simultaneous color generation systems are discussed with reference to both camera and display. Performance parameters considered include: a) Sensitivity and Signal-to-Noise b) Static and dynamic resolution c) Picture anomalies d) Reliability and Maintainability. A review of the capability of various camera and display devices suitable for field sequential color systems is included.
The limitations of displays for visual flight simulation are described and the need for a collimated display demonstrated. The results of a comprehensive investigation into methods of achieving collimated displays for flight simulation are given and the subse-quent development of the 'Duoview' display, which gives an undistorted view from all parts of the cockpit, is described and the performance indicated.
Two visual display systems for use with flight simulators, the Singer variable anamorphic motion picture system (referred to as the VAMP*) and the Singer scanned motion picture system (referred to as the SCAMP*), both use, as the image storing medium, film taken from an aircraft flying a prescribed flight path. Terrain photographed from a given point and attitude, so having a given perspective, is altered by both systems to show the same terrain viewed from a different point and attitude, with correspondingly different perspective. Data recorded during filming provide a digital computer with information about the position, velocity, etc. of the filming aircraft. Differences between the filmed flight path and the simulator flight path are used to cause the systems to correct the visual scene perspective to that corresponding to the simulator's actual spatial position. In the VAMP Visual System, the changes are produced optically, and in the SCAMP Visual System, electronically. Used with flight simulators, both systems enable pilots to "fly" maneuvers under various day and night conditions, in fog, and under, in and above clouds.
With the use of miniature scenes and models, it is possible to synthesize realistic images by combining a scene containing normal size objects with a scene containing miniature objects. The normal size scene is generally termed the foreground and the miniature is termed the background. The resulting synthesized image is generated from the foreground and background scenes using an advanced matting technique. Prior to the development of the Magicam system, camera movement during the matte had little success with holding adequate perspective match in order to achieve realistic composite images.
Measurement of resolution, modulation, and contrast on visual displays for flight simulators must be made in order to objectively assess visual system performance. A method of analyzing the image quality of a total visual system is to place a television camera at the pilot's eyepoint and set up to look at a small portion of the visual display. The camera's video output can then be analyzed to determine the level of performance. Resolution and modulation data is obtained on the entire visual system including the display device and image generation system. The observer camera must have several special qualities in order to yield accurate results. High sensitivity is needed because most visual displays have low brightness. A linear response with brightness is also required. By modifying an average-quality vidicon camera with a silicon target vidicon and a unique beam blanking circuit, a low cost camera can be used for resolution and modulation measurements.
The previous sessions of this seminar and the papers included in them have concentrated on what is generally agreed to be the most dificult outstanding simulator technical problem: the reproduction of the out-the-window visual scene. This session is designed to remind the audience that however challenging this problem is, it is the problem of one subsystem of a larger total system the simulator itself.
Northrop's capability and experience in flight simulator development utilizing closed-loop visual and motion systems has been applied in major aircraft programs for over a decade. Although our first generation of flight simulators would seem rather crude com-pared to todays standards, it was nevertheless a beginning. The development of wide angle visual display systems at Northrop will be reveiwed in terms of chronology beginning with hardware development history and leading into a discussion of current hardware and future development plans.
The US Air Force Simulator for Air-to-Air Combat (SAAC) Advanced Development Program which is currently nearing completion is a long range effort to develop a high fidelity simulator capable of simulating the total environment of one-on-one aerial combat. This paper discusses the key aspects of the visual system including multiple facet optical display, flexible high performance display electronics, image generation system, and special cathode ray display tubes.
The UCLA Driving Simulation Laboratory has recently been updated to provide for a 360 degree motion picture portrayal around a full size car. The build-up includes a rear screen which duplicated the existing front 8 ft. radius cylindrical screen and two side screens to fill the gap between front and rear scenes. The motion picture frame speed from the four projectors required to project the images is synchronized with the speed of rollers under the rear wheels of the car. A vehicle was constructed to carry four cam-eras at highway speeds with no visual obstructions in the 360 degree scene being filmed. Front and rear scenes are filmed with Vista Vision cameras equipped with "Dimension 150" optics while the side scenes are taken with 16 mm cameras.
The growing size, complexity, and operating cost of modern nuclear power plants have led to greater use of the highly sophisticated technology now available in nuclear power plant training simulators. Factors such as safety and economics strongly point to the increased use of these devices in training and requalification programs for reactor operators and key plant personnel. The number of persons needing training is increasing. Because of the industry-wide practice of base loading nuclear power stations, the percentage of these stations available for training is decreasing. Plant operators must be provided initial and refresher training in the proper handling of controls during startups, normal operations and maneuvering, shutdowns, and abnormal operations. Although training may be accomplished on the actual plant controls, the practice of taking a plant off-line for such purposes is prohibitively expensive. Futhermore, the practice of subjecting nuclear power plants to the rigors of training operations is questionable. The fact that simulators are now being built with sufficent fidelity to provide complete, continuous, real time representations of the dynamic operational conditions of nuclear power plants is particularly significant; this capability permits their use in a full range of training experiences including training in normal and emergency operating procedures.
Presenting target stimuli at the proper spacing, duration, and time interval produces movement perception in human observers. The observer sees a single target stimulus moving continuously across the physically empty space, from its first location to its second. The optimum movement perception depends mainly on such physical variables as duration of light flashes, time between flashes, intensity of flashes, and distance between flashes. The present research explores the condition under which optimum move-ment perception occurs when the distances between two stimulus flashes are 2, 5, 8, 11 and 15 centimeters. In the research average duration of flashes, p, occurred when the optimum movement perception was measured as a function of the time between flashes, t, and expressed by the rational function, μ = 1/At4 + Bt3 + Ct2 + Dt + E The results may help one to understand one of the aspects of operator performance and to aid in improving man-machine systems.
Two techniques for the computer simulation of optical and electro-optical systems have evolved: direct convolution and Fourier. These two techniques were examined from the viewpoints of computer efficiency, accuracy and ease of implementation for the purpose of developing a large scale electro-optical simulation. With the increase of the size of simulations, the trend has been toward Fourier techniques because of their computer efficiency. It was found, however, that, for the image and system being simulated, the direct convolution technique offered a significant run-time advantage. The results of that study and the development of the simulation are described.
The goal of total simulation in air carrier training programs depends on visual simulation capable of accomplishing all required landing maneuvers. Defining the visual cues that are necessary to safely maneuver and land an airplane under various weather conditions is the key factor in developing total simulation. A research program is suggested to establish basic technical information on the visual requirements for landing jet aircraft. The information produced could be useful to several segments of the aviation community.
In the broadest sense, the "Future Objectives of Simulation" must be precisely the same as the present objectives of simulation. These objectives, in research and develop-ment or engineering simulation, must be to produce research and development results which, compared to equivalent vehicle time, result in decrease in cost and increase in safety. When one of these aerospace flight simulators is used as a training device, as it some-times is before the first flight of an aircraft, there is the additional objective of increasing pilot proficiency. However, it is fair to extend the subject to include how these future objectives will be met, who will use simulators to meet them, and what the simulators might look like.
The Army is planning and preparing for more extensive use of simulation in the future to improve training effectiveness, to reduce training costs and to conserve fuel and ammunition. A variety of simulation systems are being developed for Army applications. These include gunnery and scoring systems for school and field use, conduct of fire trainers, and more sophisticated visual systems for operational flight trainers.