Testing performed on type A and type C school busses. Tested acceleration and braking performance using Vericom DAQ. Multiple drivers were used; from inexperienced to experienced bus drivers. Results show initial acceleration g spike and then settled at lower level. These are close to passenger car acceleration rates. 10% difference between all the different drivers (normal accel), 6 % difference for rapid acceleration.
Two tractors (different HP) with different loads (3700 lbs and 0). Automated transmissions, still have clutch, but only used when starting from stop. 100m acceleration test, (small uphill grade .0037 m/m). Started tractor in 1st 2nd and 3rd gears. two accelerator positions (60% and 100%). In general there was not much difference in acceleration between unloaded but significant difference between loaded. Compared these results to calculated constant acceleration. Assumptions of constant accel should be made cautiously. Vehicle HP was made little difference for unloaded vehicles.
Brake force is non-linear. Looked at relationship between PSI and brake torque. Looked at Heusser model and the variables associated with model. Details procedure to find brake force as a function of pushrod stroke. Showed relationship between static stroke and torque output. Able to calculate an adjusted drag factor from data from braking tests. Compared new model to test data. For most cases, model matches test data (not for many disabled brakes).
Tested 1995 BMW with ABS, 2003 Buell, 2005 Harley 1200. Used VBox III at 100 hz for data acquisition. Tested at speeds of 25 45 and 60. Tested 3 riders with 3 braking strategies (front, rear, both). Calculated stopping distance by integrating velocity. Calculated drag factor. Also calculated confidence interval on those results. BMW with ABS had highest drag factor. Similar results for others. At higher speeds, some motorcycles had higher drag factors at higher speeds. Analysis was done for braking differences between front, rear and both. Also performed all the same tests on a wet surface. Analyzed braking marks for the test and showed differences between braking strategies. Calculated the braking speeds from measured braking marks showed a 10% error on average. Rear brake does not always improve stopping distance.
Quantify vehicle rolling deceleration for late model vehicles, including CVT and hybrid vehicles. Studied post-impact roll (in or out of gear, automatic transmission and manual). Many classes and makes were tested. Used non-contact speed sensor (5th wheel), Vbox 100Hz and handheld GPS. Calculated deceleration values for small speed increments. Plotted deceleration over speed (coasting). CVT produces different results from manual and automatic transmissions. Average deceleration rates were presented as well.
Proposed rollover model is an extension of model proposed by Chen and Guenther previously. Model is focused on two parameters: translational velocity vs time and roll velocity vs time. Breaks down model into two phases, tumbling phase (sliding and rolling). Proposed modeling producing velocity and roll rate as a function of time consisting of trip phase airborne phase, sliding phase and rolling phase. Broke down model into a simple free body diagram. Breaks up equations into four parameters in sliding and rolling phase. Used 12 rollover tests to validate model.
Study aims to quantify error in critical speed formula. Not a challenge of the validity of the critical speed formula. Author states that small errors in the field have a large effect on calculations. Paper shows sensitivity of yaw measurements. Extensive literature review was done to analyze theory on yaw critical velocity. Author summarizes literature review and makes important notations on benefits and flaws. Author provides suggestions for documenting yaw markings. A list of recommendations, from the literature review, is compiled. Case study was also presented. Critical speed formula is a viable solution. It is only really accurate to 10 - 15%.
Analyzed previously published model for friction during rotation. Measured deceleration of a vehicle while it is rotating and measuring vehicle trajectories. Vericom and Vbox was used as data acquisition. Test conducted at the WI state patrol with a 2005 Crown Victoria. Disabled ABS and provided ability to shut off brake valves to one side of vehicle. Tested at over 50 mph, applied hard braking to one side of the vehicle. Conclusions: Simplified models in literature were validated to testing with small error.
Author reviewed classical critical speed yaw analysis. Investigates yaw marks with a decreasing radius. A yaw test was performed. The position data (x,y) was plotted. That data was attempted to be fitted as a function. Spiral was curve fit to the yaw markings. This curve fit was optimized with total least squares. Non-linear problem requires an iterative solution. The yaw markings were fit to many mathematical formulations of spiral types. Author compared the different spiral formulations to the actual test data. The author has created spreadsheets for calculating the spiral parameters at: http://tucrrc.utulsa.edu/Publications.html
Author compared multiple models for calculating speeds for yawing and braking vehicles. Also compared these models to non-braking models. Conducted some dynamics tests with a 2008 Chevy Malibu. Driver was told to brake for one test and let the vehicle come to rest in another test. Author shows the results of the tests and how the different models compare to the test data.
This paper investigates trailer side underride crashes and the associated damage. The author classifies the different types of underride vehicle damage patterns. Authors performed crash testing of high and low profile passenger vehicles impacting the side of a trailer at 60 degrees. Impact speeds ranged from 15 to 39 mph. The authors were able to derive an analytical method to derive an accurate speed estimation for low and high profile vehicles involved in trailer side underride collisions.
First objective of the study was to validate a finite element model of a 1990 Finite Element model of a Ford Taurus provided by the National Crash Analysis Center at George Washington University. Objective two was to design an analysis for a frontal collision with an underride and lateral offset. The author ran finite element simulations with the Taurus model (barrier test). They were also able to measure the crush by tracking the nodes of the Finite Element Simulation. The author performed this simulation with different crash modes (different barriers). They concluded that the relationship between the square root of crush energy and residual crush profile area is linear and invariant regardless of lateral offset or vertical offset. Some limitations include a 10% error and only to 20-40 mph.
The objective of the study was to quantify the uncertainty in CRASH3 Damage calculations. The last uncertainty evaluation was 30 years ago. Also wanted to investigate the difference in more precise crush measurements have on the calculations. Authors generated variations in CRASH3 parameters and used these variations in a monte carlo simulation to calculate uncertainty. Used NCAP and FMVSS frontal barrier tests for data source. 10,000 iterations for each scenario (Monte Carlo). Author presented a reconstruction example. Conclusions: variations in some CRASH3 parameters (d0, d1) dominate uncertainty in Ec. Range in pre-impact speed can be as high as 40 mph.
Advances in acquisition technology have greatly improved accuracy and broadened application. Author describes the basics of photogrammetry and the use of high density laser scanners to capture 3D data. The author showed an example of how to apply both of these technologies in a example crash scene. The process of using a RGB mapping process was compared to conventional methods. Conclusions: Point cloud photo projection locates evidence with accuracy comparable to PM/Total Station. Method requires no additional software. 3D DTM and data density ensure precise location on complex surfaces - no planar assumptions. 25% faster.
The authors provided a brief history of the photogrammetry methods in accident reconstruction (from the 70s to present). Some remaining challenges in this topic are fixed focal length, lens distortion, autofocus effects, incident camera unavailable. This paper describes and demonstrates a potential remedy for these challenges. Author shows how to make corrections for different lenses in different cameras. Tested seven cameras (DSLR and point-shoot) and provided four network studies: Z-D calibration, EXIF, Z-D exemplar calibration and Focal Length from One Image. The author provides experimental results (accuracies) for these variables (both large and small scale). Conclusions: Z-D Networks: Eliminate fixed focal length constraint, includes correction for lens distortion, compensates for autofocus, accuracy comparable to fixed focal length CRP, suitable accuracy for AI, etc.
Bill starts the presentation with a general history of CDR tool and how the industry contributes to the tool development. He also provides a general description of the CDR tool and components. Current coverage list was also presented. Then and now: OEMs 1 (2000) 5 (2012), estimated vehicles on road covered by CDR 35M (2000), 90M+ (2012). Reports (up to 40+ pages, 2012) have drastically changed since 2000 (5 pages). Many additions to CDR software have been added since 2000 (CSV export, VIN read, internet distribution, glossary, ETR). for 2012, there should be more cables (14), software releases, new coverage. Originally only one OEM covered by CDR, in 2012 13 total OEMs should be covered by tool (in progress). Three new OEMs should be announced this year with 7 new cables.
Author presents general history/description of GPS technology. This can be used in Accident Investigation because some GPS products log a position history. Some sources of GPS error include: placement of the antenna, number of satellites, external obstructions, reflections, and atmospheric effects. 3 GPS units were tested in this study. Driving maneuvers were performed to compare to these GPS units. Author describes the accuracy of the compared units for a single lane change, partial lane change, and a ‘no lane change’ (around a curve). Author also took a look at time lags in the recording data and the differences in speed recording between the different GPS units. They were also unable to capture subtle position changes, ie, a lane change.
Authors summarize networking principles on CAN, standardized CAN data (J1939), and unknown CAN data in passenger cars. A general history of CAN was presented. The general CAN data message process was summarized. Author shows how decoding CAN data with an oscilloscope is possible. An example was shown how to take CAN data from the scope to determine the speed on the data bus. Obtaining data can be done through many different products. Crash test was done while monitoring the CAN data traffic. Author introduces a concept of how to use CAN data to synchronize multiple measurement devices. He also discusses how the CAN data is asynchronous data and this can be assessed. More information, programming codes are available at http://tucrrc.utulsa.edu/Publications.html
Author discusses the utility and accuracy of ECM EDR data. A literature review of this data accuracy was presented. This study was focused on the Bendix ECU data information. The goal of this project was to determine which Bendix ECUs store data and to test them for accuracy. They also wanted to investigate concerns with power loss on these units. Author describes the differences between the Bendix ECUs that are presently available. Tests were done to measure the accuracy of the ECU (acceleration and braking). The data was captured by a manual trigger. In conclusion, Bendix ECU diagnostic data may be a source of collision information.
Authors performed tests where they created a non-deployment “trigger” event (steady state and ABS). CAN bus was recorded and VBOX GPS was used to measure position. The goal of the study was to compare the pre crash EDR data to the CAN bus and GPS. GPS to EDR-recorded speed difference was presented. Under steady state, EDR pre-crash speed was measured and compared. Under hard braking, the pre-crash speed differences were larger. The authors also discussed the 6th “bonus” point in the report.
The objective of this paper is to see if EDR report is accurately reporting the pre-crash data. CAN data was compared to the CDR report data and the refresh rate was investigated. Three Toyota vehicles were tested in this study. 11 test conditions were tested (33 in total) including braking, coasting, static tests, etc. Test results for these conditions were analyzed and presented. For pre crash data, tolerance bars were determined for Toyota EDR data.
Users browsing this forum: No registered users and 4 guests