One of the most important data elements recorded in the National Automotive Sampling System / Crashworthiness Data System (NASS/CDS) is the vehicle change in velocity, or ?V. ?V is the vector change in velocity experienced by a vehicle during a collision, and is widely used as a measure of collision severity in crash safety research. The ?V information in NASS/CDS is used by the U.S. National Highway Traffic Safety Administration (NHTSA) to determine research needs, regulatory priorities, design crash test procedures (e.g., test speed), and to determine countermeasure effectiveness. The WinSMASH crash reconstruction code is used to compute the ?V estimates in the NASS/CDS. However, the reconstruction accuracy of the current WinSMASH version has not previously been examined for side impacts. Given the importance of side impact crash modes and the widespread use of NASS/CDS data, an assessment of the program's reconstruction accuracy is warranted. The goal of this thesis is to quantify the accuracy of WinSMASH ?V estimations for side impact crashes, and to suggest possible means of improving side impact reconstruction accuracy. Crash tests provide a wealth of controlled crash response data against which to evaluate WinSMASH. Knowing the accuracy of WinSMASH in reconstructing crash tests, we can infer WinSMASH accuracy in reconstructing real-world side crashes. In this study, WinSMASH was compared to 70 NHTSA Moving Deformable Barrier (MDB) - to - vehicle side crash tests. Tested vehicles were primarily cars (as opposed to Light Trucks and Vans, or LTVs) from model years 1997 - 2001. For each test, the actual ?V was determined from test instrumentation and this ?V was compared to the WinSMASH-reconstructed ?V of the same test. WinSMASH was found to systemically over-predict struck vehicle resultant ?V by 12% at time of vehicle separation, and by 22% at time of maximum crush. A similar pattern was observed for the MDB ?V; WinSMASH over-predicted resultant MDB ?V by 6.6% at separation, and by 23% at maximum crush. Error in user-estimated reconstruction parameters, namely Principal Direction Of Force (PDOF) error and damage offset, was controlled for in this analysis. Analysis of the results indicates that this over-prediction of ?V is caused by over-estimation of the energy absorbed by struck vehicle damage. In turn, this ultimately stems from the vehicle stiffness parameters used by WinSMASH for this purpose. When WinSMASH was forced to use the correct amount of absorbed energy to reconstruct the crash tests, systemic over-prediction of ?V disappeared. WinSMASH accuracy when reconstructing side crash tests may be improved in two ways. First, providing WinSMASH with side stiffness parameters that are correlated to the correct amount of absorbed energy will correct the systemic over-prediction of absorbed energy when reconstructing NHTSA side crash tests. Second, providing some treatment of restitution in the reconstruction process will correct the under-prediction of ?V due to WinSMASH's assumption of zero restitution. At present, this under-prediction partially masks the over-prediction of ?V caused by over-prediction of absorbed energy. If the over-prediction of absorbed energy is corrected, proper treatment of restitution will correct much of the remaining error observed in WinSMASH reconstructions of NHTSA side crash tests. / Master of Science

Each year roughly there are roughly 40,000 traffic-related fatalities. Common roadside objects such as trees, poles, guardrails, embankments, culverts, and fences result account for roughly 46% of these fatalities. Efforts to reduce to injury risk and risk exposure in these crashes have been hampered by the difficulty in performing reconstructions. To address the need for accurate reconstructions in order to assess injury risk, a vehicle-specific stiffness database was added to the WinSmash reconstruction program. This single modification increased the average estimated delta-V by 8% and reduced error from 23% to 13% on average. A method to extend the WinSmash energy-based reconstruction approach to poles and trees that were damaged or broken was implemented to provide delta-V estimates for these crashes. The error of the pole and tree reconstruction component was roughly 44% but still represented a significant step forward for these crashes which previously could not be reconstructed. The use of strong-post w-beam guardrail along roadsides is the primary method by which exposure to risk is reduced. Efforts to model guardrails using finite element methods were hampered by the large number of unknowns and lack of knowledge about the sensitivity of the crash outcome to each variable. Through a parametric study the soil properties and rail mesh density were identified as the most significant influences in simulation outcome. This knowledge was applied to finite element models of damaged guardrail to identify when the damage compromises the guardrail ability to prevent risk exposure. Models of guardrail with rail deflection, missing posts, and missing blockouts identified rail deflection over 6 inches and any number of missing posts as hazardous conditions. The removal of a single blockout was found to be acceptable if not desirable. These findings have far-reaching implications. The enhanced WinSmash reconstruction program has been adopted by NASS/CDS to generate delta-V estimates used by researchers in all areas of transportation research. The identification of hazardous guardrail was of great interest to transportation agencies responsible for prioritizing and performing repairs of damaged guardrail. / Ph. D.

Pedal misapplication (PM) is an error in which a driver unintentionally presses the wrong pedal. When drivers mistake the accelerator pedal for the brake pedal, the vehicle experiences a sudden unintended acceleration, and the consequences can be severe. A brief history of PM is covered, and several novel studies of PM are described. The goals of these studies were as follows: 1. Identify and analyze multiple samples of PM crashes from a variety of data sources using both established and novel methods to gain new insight into the characteristics and frequency of PM crashes. 2. Use the confirmed, real-world PM crash data to develop a custom vehicle dynamics simulation and evaluate the overall potential safety benefit of a theoretical PM advanced driver assistance system. Using an established keyword search identification method and two unique crash datasets, a PM crash frequency of approximately 0.2% of all crashes was found. These PM crashes were typically rear-end or road departure crashes in moderate- to low-speed commercial or residential areas. Female drivers and elderly drivers were more often involved in these PM crashes, which generally featured slightly lower injury severities and often involved inattention or fatigue. Anecdotally, PM crash narratives contained repeated evidence of unexpected events, driver inexperience, distraction, shoe-malfunction, extreme stress, and medical conditions/emergencies. A novel PM crash identification algorithm was developed to detect PMs from time-series pre-crash data. This algorithm was applied to a sample of crashes with event data recorder data available, and a frequency of 4.3% of eligible crashes were found to have exhibited PM behavior, suggesting that PM crashes may be more prevalent than previously thought. While the data from these crashes suggested that a PM occurred, this dataset lacked sufficient data regarding driver intention, which is necessary to confirm each crash as PMs. The characteristics of these PM-like crashes were analyzed and found to be largely similar to those of previous samples, with notable exceptions for higher proportions of male drivers, higher travel speeds, and higher maximum injury severities. More robust data from a naturalistic driving study (NDS) was acquired, and the novel algorithm was applied to all of the sample’s eligible crashes. Because the NDS data contained more data elements such as driver-facing video, crashes that exhibited PM behavior were individually inspected to confirm PM. This produced a PM crash frequency of 1.1%. The characteristics of these confirmed PM crashes were investigated, but a small sample size limits the generalizability of the results. Lastly, crash data from confirmed, real-world PM crashes was used to inform a custom vehicle dynamics model into which a theoretical PM advanced driver assistance system was simulated. The effect of the accelerator suppression system on crash avoidance and mitigation was evaluated to assess its potential safety benefit, which was found to be highly dependent on system threshold values and largely underwhelming in the absence of supplemental braking. The results indicated that a system that detected PM, suppressed acceleration, and applied braking could provide a substantially higher safety benefit. / M.S. / Pedal misapplication (PM) occurs when a driver presses the wrong pedal. When drivers mistake the accelerator pedal for the brake pedal, the vehicle experiences a sudden unintended acceleration, and the consequences can be severe. A history of the controversial subject of PM is covered, and several novel studies of PM are described. In these studies, PM crashes are identified among documented real-world crashes. This is done in three phases: (1) using narratives written by law-enforcement officers or crash investigators, (2) using event data recorders, or “black boxes,” that store vehicle data prior to crashes, and (3) using naturalistic driving study data, including video recordings of subjects during daily driving. These data are analyzed to develop the understanding of how often PM crashes occur and what factors are common among them. It is discovered that the frequency of PM crashes may be an order of magnitude greater than previously estimated. In the final study, real-world PM crash data is used to virtually reconstruct PM crashes and apply an advanced driver assistance system designed to detect PM, suppress the accelerator input, and reduce the severity of the crash or prevent it altogether. By simulating a wide range of system variations, we develop a sense of the feasibility of such a system’s implementation and overall safety benefit.