Crash reconstruction of pedestrian accidents using optimization techniques

https://doi.org/10.1016/j.ijimpeng.2008.01.012Get rights and content

Abstract

Numerical simulations of vehicle-to-pedestrian crash (VPC) are frequently used to develop a detailed understanding of how pedestrian injuries relate to documented vehicle damage. Given the complexity of the event, modeling the interactions typically involves subjective evaluations of the pre-impact conditions using a limited number of simulations. The goal of this study is to develop a robust methodology for obtaining the pre-impact pedestrian posture and vehicle speed utilizing multi-body simulations and optimization techniques. First, a continuous sequence of the pedestrian gait based on the literature data and simulations was developed for use as a design parameter during the optimization process. Then, the robustness and efficiency of three optimization algorithms were evaluated in a mock (idealized) crash reconstruction. The pre-impact parameters of the pedestrian and the vehicle models were treated as unknown design variables for the purpose of validating the optimization technique. While all algorithms found solutions in close vicinity of the exact solution, a genetic algorithm exhibited the fastest convergence. The response surfaces of the objective function showed higher sensitivities to the pedestrian posture and its relative position with respect to the vehicle than to the vehicle speed for the chosen design space. After validating the methodology with the mock reconstruction, a real-world vehicle-to-pedestrian accident was reconstructed using the data obtained from the field investigation and the optimization methodology. A set of pedestrian and vehicle initial conditions capable of matching all observed contact points was determined. Based on the mock and real-world reconstructions, this study indicates that numerical simulations coupled with optimization algorithms can be used to predict pedestrian and vehicle pre-impact conditions.

Introduction

Pedestrian fatalities in traffic accidents continue to be a serious and costly social problem. According to international statistics, pedestrian fatalities represent 65% of the road crash related fatalities in the world [1]. While the highest percentages are recorded in developing countries, pedestrian fatalities still represent 13–27% of road fatalities in the US, the European Union, and Japan. To address this serious public health problem, researchers have been developing physical and computational pedestrian dummies to further understand pedestrian injury mechanisms and to develop appropriate injury countermeasures. Recently, multi-body system (MBS) simulations of vehicle-to-pedestrian crashes (VPC) have been used to reconstruct real-world pedestrian accidents based on information documented in pedestrian databases [2], [3], [4], [5]. The reconstructed pedestrian kinematics data obtained in these studies have been used as initial conditions in pedestrian subsystem tests [3], [4] or finite element (FE) simulations [5] to correlate measured and calculated injury parameters in pedestrian models with the presence of injury in pedestrian accidents. Since field investigations of pedestrian accidents estimate the impact vehicle speed within ranges [6] and usually provide little information about the pedestrian stance, the typical approach for modeling these events involves subjective evaluations of the pre-impact conditions using a limited number of simulations. While many previous studies (cf., [5], [7]) have found pedestrian kinematics and injury outcomes to be significantly affected by the initial posture of pedestrian, a limited number of stances (e.g., the six stance sequence proposed by [8]) have been used in previous crash reconstructions. Furthermore, a trial-and-error approach can be time consuming and may not lead to the best accident scenario since it is highly dependent on the analyst's experience and intuition.

Given the shortcomings of traditional crash reconstruction approaches for pedestrian crash reconstruction, the current study has two goals. The first goal is to implement a continuous sequence of pedestrian gait based on data from the literature to model all the stages within the pedestrian walking gait prior to the impact. Given an established gait relationship, the second goal was to develop an improved methodology for identification of pre-impact vehicle speed and pedestrian posture in pedestrian accidents using rigid-body simulations and design optimizations techniques.

Section snippets

Numerical simulation of vehicle–pedestrian lateral impact

Statistical studies of pedestrian crashes have shown that most pedestrians struck by vehicles are in the process of walking, as opposed to standing stationary, prior to the crash. The Pedestrian Crash Data Study (PCDS) [9] containing US data found that 55% of pedestrians were walking prior to the crash and that 72% of pedestrians were struck on their lateral sides [10]. A similar study examining German data [11] found that 56.1% of the pedestrians struck by vehicles were walking, and in 81.1%

Pedestrian in walking stances

The stances of the pedestrian model obtained using functions of joint angles and H-point corresponding to 10 gait parameters (Section 2.1.1) are provided in Fig. 12. It can be observed that the gait cycle of the right lower limb is divided into two phases: a stance phase and a swing phase. The stance phase is the entire period when the right foot is on the ground (about 60% of the gait cycle) and is divided into two double limb stances (period when both feet are on the ground) connected by a

Discussion

Previous studies [7], [34], [35] showed that pedestrian pre-impact posture has a significant influence on pedestrian kinematics during vehicle-to-pedestrian simulations. To better represent these pre-impact parameters, this study implemented a continuous gait sequence of the Madymo pedestrian model [13] based on literature data and simulations. Although the angles of the lower limb joints were defined according to the test data recorded on a subject with anthropometric characteristics close to

Conclusion

The main goal of this paper was to investigate the application of optimization techniques to the reconstructions of pedestrian crashes. It was shown in a mock or “ideal”, crash reconstruction problem that optimization algorithms combined with an appropriate objective function have capability to identify accurately the pre-impact conditions of pedestrian and vehicle. While all algorithms found solutions in close vicinity of the exact solution, the genetic algorithm exhibited the fastest

References (40)

  • Mizuno Y. Summary of IHRA pedestrian safety WG activities – proposed tests methods to evaluate pedestrian protection...
  • Pedestrian crash data study(PCDS), data collection, coding and editing manual

    (1996)
  • A. Chidester et al.

    Final report – the pedestrian crash data study

    Proceedings of the 17th int conf on the enhanced safety of vehicles

    (2001)
  • M. Kuehn et al.

    An advanced testing procedure for pedestrian–car collisions

    Int J Vehicle Safety

    (2005)
  • J. van Hoof et al.

    Improving pedestrian safety using numerical human models

    Stapp Car Crash J

    (2003)
  • MADYMO human models manual

    (2006)
  • RAMSIS manual

    (1997)
  • Schneider LW, Robbins DH, Pflug MA, Snyder RG. Development of anthropometrically based design specifications for an...
  • H. Ishikawa et al.

    Computer simulation of impact response of the human body in car–pedestrian accidents

    Proceedings of the 37th stapp car crash conference

    (1993)
  • J.K. Yang et al.

    A human-body 3D mathematical model for simulation of car–pedestrian impacts

    Int J Crash Prev Injury Contr

    (2000)
  • Cited by (0)

    1

    Tel.: +81 28 677 7594; fax: +81 28 677 7230.

    View full text