Analysis of motion is an important aspect of automotive safety crash testing. As in other applications, high speed photography is frequeutly the best way to obtain motion analysis, and it is usually important to now the accuracy of the motion data that is obtained. In any high speed film analysis system there can be many sources of data error. Generally, error results from either optical distortion, and resolution limits or from limitations in the setup of the test being photographed. Methods of analysis for error of this latter type, such as parallax error, are quite dependent on the specifics of a given test setup and are not discussed in this paper. Techniques for analysis of distortion and resolution error can be quite universal. The following is intended to provide techniques for quantifying in a useable fashion, the error from distortion and resolution limits.

One of the responsibilities of the Safety Research and Development Laboratory at the Milford Proving Ground is to conduct full scale impact tests. Nearly all of the testing is performed at a facility whose central feature is a standard SAE barrier. The barrier is located within a building, and is approached on a covered test roadway. (Figure 1).

A photographic method was developed to meet the need for measuring the low level luminance distribution of highway and obstacle target arrangements in a program of studies on automobile headlighting (Ref. 1).

Calspan Corporation of Buffalo, New York is currently developing a computerized vehicle to aid in the investigation and reconstruction of automobile accidents. This vehicle includes an optical measuring technique which will enhance the speed and accuracy with which accident evidence is collected at the scene. A unique feature of the system is the ability to scientifically reconstruct the accident, based on the optically gathered evidence. It is felt that this system, which improves the accuracy and consistency of accident reporting, will have a positive influence on efforts to improve traffic safety on our nation's highways.

To improve highway safety, significant research is being conducted on new and improved automotive restraint systems by both Government and industry.

One would think that since we have had automobiles on the road for over 70 years we would, by now, completely understand the role vision plays in driving. This is certainly not the case. Despite some excellent research in this field, we have just begun to formulate the questions which need asking. One cannot separate the optics of the vehicle from the driver, especially in the dynamic driving environment. The direct vision out of a vehicle should be such that the driver can acquire all the visual information necessary to safely perform his task and, at the same time, the quality of this information should be such that it is very unlikely that it be misinterpreted.

The determination of the recognition of visual targets in a driver's forward field of view which is primarily foveal vision is rather straight forward. The driver sights an object, recognizes it as an obstacle to be avoided, a passive object, or something which provides information. A number of studies have determined the scan patterns of drivers, how he sights objects during the day and night and how effectively he can use the visual input to help him avoid obstacles and remain in his lane. SAE Recommended Practices and Federal Motor Vehicle Safety Standards are considered when determining the area of forward visibility to be kept unobstructed in order for the driver to view important forward targets.

Studies conducted at a number of locations around the country by Southwest Research Institute for the U.S. Department of Transportation have shown that glare from vehicle headlights at night has a much greater impact on motorists from a psychological standpoint than it does from a physiological, vision-degrading standpoint.

The quality of motorist information systems is of considerable importance to the safety and efficiency of our Nation's highways. Accordingly, the Federal Highway Administration continues to be engaged in a number of activities that pertain to improving highway signing. For example, real-time information systems are now the subject of intensive research and development efforts by the Office of Research. Modern technology now permits communication of real-time information to drivers by means of variable message displays. Thus motorists can be given advanced warning of freeway incidents or other prevailing conditions such as fog or icy roads. In other words, real-time variable message signs provide the capability of giving the driver timely and useful information which corresponds to changing traffic conditions. Examples of major types of such signs are rotating drums, matrix lights, matrix discs, rolling screens, blank out incandescent, neon and fluorescent lamps signs, and new fibre optics signs.

Hign-speed x-ray cineradiograpny application to biomechanics and impact trauma research is a recent development. There is growing interest because of the need to observe directly, quantify, and analyze the mechanisms of these phenomena in a precise and accurate way. Advances in x-ray sources and detectors directed toward high frame rates have occurred in response to these needs.

The human legs provide strong resistance to motion during a car crash. Laboratory simulators of large deceleration accidents typically utilize anthropomorphic dummies in place of human volunteers. In order to better understand the biomechanics of car crash injury and to develop more realistic dummies, improved knowledge of knee kinematics must be obtained. Most mathematical and physical analogs of human knees consider the motion to be modeled by a hinge or pin type joint. In recent years, many complex models have been suggested to better approximate the knee in walking. The present research suggests that the simplicity of a pinned joint be retained, but that the pin location be chosen to best simulate the in-plane motion of the human knee. High speed flash x-rays of the knee region were photographed during a highly simplified mock car crash. The relative positions of a rotating femur about a fixed tibia were observed over a 25° range or rotation. Using an optimization method, an effective pivot location was found that most closely matched the complex motion of the actual knee. The results of this preliminary study show this optimum pin position to be nearly 2.5 cm from the pin location that would have been obtained by constraining the location to lie along the tibia centerline (a geometric center). Also, the appropriate pin position was located more anterior than for tests with non-weight bearing and cadaver subjects.

This study describes a numerical method of curve fitting with provisions for subsequent differentiations. The method has been used to provide a basis for the analysis of high-speed movies of human volunteers on the Daisy decelerator at Holloman Air Force Base. Since deceleration measurements were not available from these classical studies of human tolerance, significant biomechanical tolerance and whole-body response information are only available from the analysis of the high-speed movies.

Sophistication in crash simulation testing has lead to a need for exacting techniques to locate the test subject relative to the impact sled. A laser system is described that facilitates the alignment of a sagittal plane of the subject and the positioning of target points on the subject to specific sagittal plane coordinates relative to the sled. The system is compatible with anthropomorphic dummy, human cadaver, and volunteer subjects.

The Naval Aerospace Medical Research Laboratory Detachment cinephotographic system in use in an investigation of human and human analogue dynamic response to impact acceleration is intended to provide documentary film footage as well as accurate measures of the kinematics of these responses in two and three dimensions. A complete description of this system is undertaken in both its hard and software phases. Accounts of filming and lighting schemes for "G" environments, digitization of photographic data, and a mathematical solution for the position and orientation of a multiply targeted rigid body as acquired by a number of fixed cameras are included in the development.

Let us confess at the very beginning of this article that Lord Rayleigh did not have very much to do with automotive safety. In fact, when he was born in 1842 the entire concept of self-powered vehicles was still considered to be science fiction, although the steam locomotive had just been invented and railroads were beginning to spread outwards from London. Towards the end of his life, the family acquired a touring car to take him to the railroad station for his many trips to London, but I do not believe Rayleigh himself ever learned to drive a car.

Two drivers drove an automobile over a 15 mile two-lane course in the daytime and at night using American and European low beam headlamps and a mid beam. The eye fixations of the drivers were measured, using a silicon diode TV camera and light reflection from the cornea as the recording device. Comparisons between the eye fixations used in daytime and at night, suggested that dwell time was longer when looking straight ahead at night than in the daytime, and there was a reduction in the proportion of the viewing time devoted to the left lane at night than in the day, when there was no oncoming vehicle. Drivers looked at approaching vehicles in both day and night conditions with glance durations of intermediate length, which increased in frequency as the separation distance between the vehicles decreased. At night preview distances were less than in the day. The characteristic shift of the eye fixations in the direction taken by the road was found in this study, in both day and night driving. At night the eye fixations were influenced by the characteristics of the beam pattern being used. On left curves the American and European low beams provided eye fixations which most closely resembled those used in daytime, whereas on straight sections and particularly on right curves the mid beam provided the most compatible distribution of glances. A revised composite mid beam, incorporating some characteristics of the European low beam was suggested as an improved meeting beam.

A method of non-destructively detecting internal anomalies or void areas in tires can be a valuable tool for tire structural analysis. Such a system has become commerically available. This system uses holographic interferometry in conjunction with vacuum stressing to non-destructively inspect tires for internal voids or anom alies. A holographic non-destructive testing (HNDT) system has been purchased by General Motors Proving Ground and has been utilized by the Tire and Wheel Engineering Laboratory for structural analysis of tires prior to, during, and after testing. The HNDT equipment is described and the process and utility of passenger car tire analysis explained.

The ACTA (Automatic Computerized Transverse Axial) tomographic scanner is designed for the radiological delineation of tissue abnormalities through the whole body, accurately portraying in pictorial and quantitative form both the nature of the lesion and its precise three-dimensional location in the body. The nature of the lesion is identified by means of the relative absorption coefficient (for X-rays), and the location of the lesion is given precisely by its position in the picture and location of its cross section. The concept of computerized transverse axial tomography involves the production of X-ray absorption profiles made at different angles in the same cross-sectional plane. A highly columnated X-ray beam scans the cross section in two composite modes of motion - a translation, or scan-pass, and a rotation. The pencil-thin beam of X-rays passes through the body and is detected by a sodium iodide crystal, whose scintillation is measured by a photomultiplier tube. The X-ray profiles are sampled, digitized, and fed into a computer, which synthesizes them into a picture giving the relative absorption coefficients of the body cross section in the plane of the scan. The cross section picture can have as many as 160 X 160 elements, each representing 1.5 mm x 1.5 mm of cross-sectional area. Two adjacent cross sections can be scanned each time (about 4 1/2 min.), each of which has a thickness of 7.5 mm and with a 3 mm distance between them. The picture is reconstructed immediately upon completion of the scan and is instantaneously displayed on a color TV console and two black-and-white consoles. The color TV display shows at a glance the areas of different ranges of absorption coefficients as areas of different colors which, in turn, will facilitate the clinical diagnosis. The picture can also be stored for permanent record.

In the field of Automotive Safety there is apparently an absence of a solution to the problem of sensing spacial intensity data in a rapidly scanned field of view and processing this data in real time. DBA Systems, Inc. has developed and is fabricating a system that will, under program control, direct a non-storage photo-emissive device beam to a discrete X, Y address location. Intensity information at and around this point is then transferred to memory for real time processing. The Automatic Programmable Film Reader (APFR) presents itself as a highly useful system when applied to the field of Automotive Safety. The APFR locates up to 30 points of interest on a single frame of film. The system then tracks and stores the new locations of these points in the ensuing frames. Tracking, locating, and advancing of film is accomplished fully automatically under program control and does so with an accuracy of less than 10 microns. The total time required to track and store the locations of the full complement of points per frame is 10 seconds. The ability to track points of interest within a given field of view in real time with a non-storage camera is further enhanced by a choice of optics and projection heads enabling the operator to observe 16mm, 35mm or 4" x 5" cut film records. In addition to the storage of location information on each frame, the frame itself is magnified and projected onto a viewing screen for visual observance by the operator. This paper describes the solution to the problems of inaccurate and inefficient film analysis. The solution lies in the replacement of human "eyes" with an optical sensor capable of digitizing information and transferring the information at a high throughput rate to a computer for real time processing.