Modeling spatial trajectories in dynamics testing using basis splines: application to tracking human volunteers in low-speed frontal impacts

Abstract

Designing motor vehicle safety systems requires knowledge of whole body kinematics during dynamic loading for occupants of varying size and age, often obtained from sled tests with postmortem human subjects and human volunteers. Recently, we reported pediatric and adult responses in low-speed (<4 g) automotive-like impacts, noting reductions in maximum excursion with increasing age. Since the time-based trajectory shape is also relevant for restraint design, this study quantified the time-series trajectories using basis splines and developed a statistical model for predicting trajectories as a function of body dimension or age. Previously collected trajectories of the head, spine, and pelvis were modeled using cubic basis splines with eight control points. A principal component analysis was conducted on the control points and related to erect seated height using a linear regression model. The resulting statistical model quantified how trajectories became shorter and flatter with increasing body size, corresponding to the validation data-set. Trajectories were then predicted for erect seated heights corresponding to pediatric and adult anthropomorphic test devices (ATDs), thus generating performance criteria for the ATDs based on human response. This statistical model can be used to predict trajectories for a subject of specified anthropometry and utilized in subject-specific computational models of occupant response.

Device-measured physical activity data for classification of patients with ventricular arrhythmia events: A pilot investigation

“…There are several approaches that investigators have applied to extract features from a time series in order to use the data to model a given outcome. While some approaches, such as using frequency-domain analysis[ 9 , 21 ] or B-splines[ 22 , 23 ], have been successful in situations where a given timeframe is meaningful, in others, such as comparing physical activity trends across a population, they do not seem to have as much relevance.…”