Search Results

The statistical analysis of vehicle crash accident data is generally problematic. Data from commonly used sources is almost never without error and complete. Consequently, many analyses are contaminated with modeling and system identification errors. In some cases the effect of influential factors such as crash severity (the most significant component being speed) driver behavior prior to the crash, etc. on vehicle and occupant outcome is not adequately addressed. The speed that the vehicle is traveling at the initiation of a crash is a significant contributor to occupant risk. Not incorporating it may make an accident analysis irrelevant; however, despite its importance this information is not included in many of the commonly used crash data bases, such as the Fatality Analysis Reporting System (FARS). Missing speed information can result in potential errors propagating throughout the analysis, unless a method is developed to account for the missing information.

The effects of vehicle “stiffness” and mass on the occupant response during a crash may be determined by evaluation of accident data. However, “stiffness” and mass may be correlated, making it difficult to separate their effects. In addition, a single-valued “stiffness”, although well defined for linear case, is not well defined for non-linear systems, such as in vehicle crash, making the separation task even more difficult. One approach to addressing the lack of a clear definition of stiffness is to use multiple definitions. Each stiffness definition can then be correlated with mass to look for trends. In this study, such an approach was taken, and the different stiffness definitions were given and their values were obtained from rigid barrier crash test data. No clear relationship between mass and stiffness appears to exist. All the stiffness measures reviewed show, at best, only a weak correlation with mass. A stiffness analysis among different vehicle types was also carried out.

A data processing algorithm is presented for determining the spatial orientation and position of a rigid body in impact tests based on an instrumentation scheme consisting of a triaxial angular rate sensor and a trialaxial linear accelerometer. The algorithm adopts the unit quaternion as the main parameterized representation of the spatial orientation, and calculates its time history by solving an ordinary differential equation with the angular rate sensor reading as the input. Two supplemental representations, the Euler angles and the direction cosine matrix, are also used in this work, which provide an intuitive description of the orientation, and convenience in transforming the linear accelerometer output in the instrumentation frame to the global frame. The algorithm has been implemented as a computer program, and a set of example impact tests are included to demonstrate its application.

A new approach for improving estimates of the kinematic response of ATDs (anthropomorphic test devices) to vehicle crash events has been developed. This approach employs the Kalman Filter; a state model based estimation approach that has been widely applied to system dynamics problems ranging from navigation to missile guidance. The Kalman Filter approach combines measurements of crash event phenomena (acceleration and displacement), kinematic models of ATD behavior and statistics of sensor noise to create precise estimates of ATD motion during a crash. This paper presents an implementation of a state model and Kalman Filter for a sensor data collected from the chest of an ATD during an out-of-position airbag deployment test. Favorable comparisons are made between the Kalman Filter model approach and traditional methods involving numerical integration and differentiation.

This paper discusses techniques developed and used by the Biosciences Group at the University of Michigan Transportation Research Institute (UMTRI) for measuring three-dimensional head motion, skull bone strain, epidural pressure, and internal brain motion of repressurized cadavers and Rhesus monkeys during head impact. In the experimental design, a stationary test subject is struck by a guided moving impactor of 10 kg (monkeys) and 25 or 65 kg (cadavers). The impactor striking surface is fitted with padding to vary the contact force-time characteristics. The experimental technique uses a nine-accelerometer system rigidly affixed to the skull to measure head motion, transducers placed at specific points below the skull to record epidural pressure, repressurization of both the vascular and cerebrospinal systems, and high-speed cineradiography (at 1000 frames per second) of radiopaque targets.

A test series using unembalmed cadavers was conducted to investigate thoracic response differences in lateral impacts between high energy (rib fractures produced) and low energy (no rib fractures produced) testing and also the response to low energy impacts for different impact directions (frontal, 45°, and lateral). Five of the test subjects were instrumented with a nine-accelerometer package and an eighteen-accelerometer array to measure thoracic response. Seven of the test subjects were instrumented with a triaxial accelerometer on the head and a six-accelerometer array to measure thoracic response. Impact events were performed with either the UMTRI pendulum impact device or the UMTRI pneumatic impact device. The subject was struck with a free-traveling mass (25 or 56 kg) which was fitted with either a 15 cm round or 20 cm square rigid metal surface.

One method of understanding the general mechanical response of a complex system such as a vehicle, a human surrogate, a bridge, a boat, a plane, etc., is to subject it to an input, such as an impact, and obtain the response time-histories. The responses can be accelerations, velocities, strains, etc. In general, when experiments of this type are run the responses are contaminated by sample-to-sample variation, test-to-test variability, random noise, instrumentation noise, and noise from unknown sources. One common method of addressing the noise in the system to obtain the underlying response is to run multiple tests on different samples that represent the same system and add them together obtaining an average. This functionally reduces the random noise. However, if the fundamental response of each sample is not the same, then it is not altogether clear what the average represents. It may not capture the underlying physics.

Reliable testing of a mechanical system requires the procedures used for the evaluation to be repeatable and reproducible. However, it is never possible to exactly repeat or reproduce the tests that are used for evaluation. To overcome this limitation, a statistical evaluation procedure can generally be used. However, most of the statistical procedures use scalar values as input without the ability to handle vectors or time-histories. To overcome these limitations, two numerical/statistical methods for determining if the impact time-history response of a mechanical system is repeatable or reproducible are evaluated and elaborated upon. Such a system could be a vehicle, a biological human surrogate, an Anthropometric Test Device (ATD or dummy), etc. The responses could be sets of time-histories of accelerations, forces, moments, etc., of a component or of the system. The example system evaluated is the BioRID II rear impact dummy.

Simple rigid body dynamics models are created to analyze the curb- and soil-trip types of rollover events and experimental methods that are used to simulate these events. Equations for the models are given, and they are integrated numerically to obtain the solution. Solutions of the models provide a break down of the energy during these events, which exposes the importance of energy absorption, unloading, and friction during the impact-and-roll process. Furthermore, the models are used to derive the critical sliding velocity under different test parameters. They are also used to understand near-critical state responses of the vehicle, and the corresponding characteristics of the signals in the phase space.

Head-to-knee interaction of the right front passenger dummy can occur in some 30-35 MPH crash barrier tests. The biofidelity and significance of these interactions as related to predicting human response was addressed in this study. In a series of laboratory experiments an instrumented headform was dropped on the dummy knee to simulate the barrier interactions. These test results were then related to the human by dropping the same headform on the cadaver leg. The instrumented headform was dropped from three heights to impact the Part 572 dummy knee at three velocities. Two impact sites and two impact angles were used. These test parameters bracketed the barrier conditions. Measurements from headform accelerometers permitted calculation of HIC value for comparison to barrier values. Comparable experiments were subsequently performed with three unembalmed cadaver subjects using the same headform and test procedures.

A comparison of the NHTSA advanced dummy and the Hybrid III is presented in this paper based on their performance in repeated sled tests under 3 different restraint systems. The restraint systems considered are: the airbag alone, the 3-point belt alone, and a combined use of the airbag and the 3-point belt. Various time-histories pertaining to accelerations, angular velocities, deflections and forces have been compared between the two dummies in order to study their repeatability. The Hybrid III appears to be more repeatable than the NHTSA advanced dummy in its response in one case, that of restraint with the 3-point belt alone. The response of the NHTSA advanced dummy in other two restraint modes, the airbag alone and the combination of 3-point belt and airbag, appears to be no less repeatable than that of Hybrid III in this series of tests.

The object of this paper is to highlight the potential limitations of numerical procedures and the need to capture the relevant physics in the FEA models for head impact studies. This is accomplished through a discussion on stress update objectivity, which assumes particular importance because it affects the accuracy of stress and strain calculations when large displacements associated with rotations, as seen in head impacts, are involved. Inaccurate stress and strain results will also result due to material rotation if the objectivity is not maintained.

Supplemental airbag safety restraint systems are an integral part of today's vehicle package. This inflatable restraint technology relies heavily on woven fabrics and particularly on knowledge pertaining to a fabric's permeability as a function of pressure drop, inflation temperature of the gas and fabric weave. While fabric permeability can be quantified by actual experimental measurements, the number and non-linearity of the variables involved make the experiments time and cost intensive. Moreover, interpolations within a given data set yield questionable results. For these reasons a feed-forward artificial neural network (ANN) technique was utilized to predict fabric permeability. This is an interpretive procedure. An ANN routine must first be trained. During this training the ANN is introduced to actual cause and effect patterns with adjustments being made by changes in weighting factors until the errors in the output variables are minimized.

Multiple axial knee impacts and/or a single lateral pelvis impact were performed on a total of 19 cadavers. The impacting surface was padded with various materials to produce different force-time and load distribution characteristics. Impact load and skeletal acceleration data are presented as functions of both time and frequency in the form of mechanical impedance. Injury descriptions based on gross autopsy are given. The kinematic response of the pelvis during and after impact is presented to indicate the similarities and differences in response of the pelvis for various load levels. While the impact response data cannot prescribe a specific tolerance level for the pelvis, they do indicate variables which must be considered and some potential problems in developing an accurate injury criterion.

This paper reports a study undertaken by the Crash Safety Working Group (CSWG) of the United States Council for Automotive Research (USCAR) to determine generic acceleration pulses for testing and evaluating advanced batteries subjected to inertial loading for application in electric passenger vehicles. These pulses were based on characterizing vehicle acceleration time histories from standard laboratory vehicle crash tests. Crash tested passenger vehicles in the United States vehicle fleet of the model years 2005-2009 were used in this study. Crash test data, in terms of acceleration time histories, were collected from various crash modes conducted by the National Highway Traffic Safety Administration (NHTSA) during their New Car Assessment Program (NCAP) and Federal Motor Vehicle Safety Standards (FMVSS) evaluations, and the Insurance Institute for Highway Safety (IIHS).

The response of the head to impact was investigated using live anesthetized and postmortem Rhesus monkeys and repressurized cadavers. The stationary test subject was struck by a guided moving impactor of 10 kg for monkeys; 25 or 65 kg for cadavers. The impactor striking surface was fitted with padding to vary the contact force-time characteristics. The experimental technique used a nine-accelerometer system rigidly mounted on the head to measure head motion, transducers placed at specific points below the skull to record epidural pressure, repressurization of both the vascular and cerebral spinal systems of the cadaver model, and high-speed cineradiography (at 400 or 1000 frames per second) of selected test subjects. The results of the tests demonstrate the potential importance of skull deformation and angular acceleration on the injury produced in the live Rhesus and the damage produced in both the post-mortem Rhesus and the cadaver as a result of impact.

Two series of impacts to the head in the superior-inferior direction using 19 unembalmed cadavers are reported. The first series of five tests was aimed at generating kinematic and dynamic response to sub-injurious impacts for the purpose of defining the mechanical characteristics of the undamaged head-neck-spine system in the S-I direction. The second series of fourteen tests was intended to define injury tolerance levels for a selected subject configuration. A 10-kg impactor was used to deliver the impact to the crown at a nominal velocity of 8 m/s for the first series, and between 7 and 11 m/s for the second series. Measurements made in the first series include the impact velocity, force, and energy, the head three-dimensional kinematics, forces and moments at the occipital condyles, and accelerations of the T1, T6, and T12 vertebrae. Impact impedance curves were also generated.

The response of the head to impact in the posterior-to-anterior direction was investigated with live anesthetized and post-mortem primates.* The purpose of the project was to relate animal test results to previous head impact tests conducted with cadavers (reported at the 21st Stapp Car Crash Conference (1),** and to study the differences between the living and post-mortem state in terms of mechanical response. The three-dimensional motion of the head, during and after impact, was derived from experimental measurements and expressed as kinematic quantities in various reference frames. Comparison of kinematic quantities between subjects is normally done by referring the results to a standard anatomical reference frame, or to a predefined laboratory reference frame. This paper uses an additional method for describing the kinematics of head motion through the use of Frenet-Serret frame fields.

Forward Collision Warning (FCW) is a system intended to warn the driver in order to reduce the number of rear end collisions or reduce the severity of collisions. However, it has the potential to generate driver annoyances and unintended consequences due to high ineffectual (false or unnecessary) alarms with a corresponding reduction in the total system effectiveness. The ineffectual alarm rate is known to be closely associated with the “time to issue warning.” This results in a conflicting set of requirements. The earlier the time the warning is issued, the greater probability of reducing the severity of the impact or eliminating it. However, with an earlier warning time there is a greater chance of ineffectual warning, which could result in significant annoyance, frequent complaints and the driver's disengagement of the FCW. Disengaging the FCW eliminates its potential benefits.

The objective of this study is to analytically model the fatality risk in frontal vehicle-to-vehicle crashes of the current vehicle fleet, and its sensitivity to vehicle mass change. A model is built upon an empirical risk ratio-mass ratio relationship from field data and a theoretical mass ratio-velocity change ratio relationship dictated by conservation of momentum. The fatality risk of each vehicle is averaged over the closing velocity distribution to arrive at the mean fatality risks. The risks of the two vehicles are summed and averaged over all possible crash partners to find the societal mean fatality risk associated with a subject vehicle of a given mass from a fleet specified by a mass distribution function. Based on risk exponent and mass distribution from a recent fleet, the subject vehicle mean fatality risk is shown to increase, while at the same time that for the partner vehicles decreases, as the mass of the subject vehicle decreases.