This paper continues a previous study of the effects of uncertainty of measurement upon accident reconstruction. The task is to identify, given the many inevitable errors of observation, the few of greatest import, so that these errors may be reduced, and to document the accuracy of the associated reconstruction.
Until recently, it was not for lack of method that such studies could not be properly performed, but for lack of good data on uncertainty of measurement. The essential data was provided in 2002 in a report by Bartlett and others of juried studies performed by volunteer field investigators, summarized and supplemented in 2003 by Bartlett and Fonda in the form of a single table of all likely errors of measurement (furnished again here). In that paper, Finite Difference Analysis (FDA) was reviewed and with the aid of the new data was applied to automotive accident reconstruction. FDA includes identification of “the vital few among the trivial many” (a Pareto analysis) as a guide to the benefits of investigative efforts both past and pending.
Both studies consider the effect of three levels of measurement uncertainty on the results of reconstruction, but previous only the speeds of approach to eccentric intersection impact were treated, by means of CRASH3 only. The present paper treats reconstruction of more output parameters of more cases by means of more algorithms, and for those cases and parameters reports as well the methodological errors of those differing treatments.
In addition to approach speed V0, which correlates with accident causation, the outputs now include values, independently based on momentum and on energy, of the speed change ▵V, which correlates with injury causation. The cases now include three more types of impact, namely symmetric intersection impact, axial (in-lane) impact, and impact with a stationary object. As before the study uses a version of CRASH3 which provides routines dedicated to FDA and optionally treats motions and forces occurring during impact.
The broadened study extends the available guidelines for the optimization of field investigation expenditures and efforts, and for the recitation of the bounds of uncertainty of result due to the uncertainties of measurement and subsequent treatment.