Skip to main content

Pedestrian pre-crash reactions and their effects on crash outcomes

Final Report Summary - PEDPCREACT (Pedestrian pre-crash reactions and their effects on crash outcomes)

Existing pedestrian protection systems do not account for the pedestrian ability to react prior to an accident situation. Under a perceived risk or an emergency situation, such as an imminent pedestrian accident, basic human survival instincts may result in sudden burst of involuntary reactions. Consequences arising from these pre-crash reactions may eventually undermine the benefits of safety systems during pedestrian accidents. Thus, the objective of the present research work was to observe and quantify pedestrian reactions just prior to an unavoidable imminent accident and then to investigate how these reactions can affect the crash outcomes.

A protocol for volunteer experiments was therefore developed and implemented. A non-invasive methodology was adapted to simulate an accident situation. A pedestrian street crossing scenario with a simulated two-way traffic flow was created using image generation and projection techniques. Volunteers equipped with 46 vicon markers and 12 wireless electromyography (EMG) electrodes were employed in this semi-virtual reality environment. Volunteers were surprised by simulated accident at a specific time while they were walking in the street crossing scenario. Vicon marker trajectories and EMG data were utilised to calculate volunteer postural movements and muscle activities, respectively. Results obtained from the volunteer experiments were subsequently used in multi-body simulations of car to pedestrian impacts to adjust the pedestrian pre-crash conditions and their effects on the crash outcomes were then investigated.

Twenty three volunteers from two age groups (12 young between 18 - 30 years old and 11 elderly between 60 - 75 years old) were recruited and 51 simulated accident trials were analysed. Experimental results revealed that both elderly and young volunteers reacted in three different strategies which were grouped here as:

(i) accelerate (when a volunteer ran, sprinted or jumped),
(ii) freeze (if a volunteer decelerated or stopped in fright) and
(iii) back-out (when a volunteer stepped back).

While majority of youngsters accelerated (60 %), elderly froze in fright (38 %) as often as they ran (40 %). Youngsters (average reaction time 501 ms) were 1.5 times faster in taking an action than elderly (average reaction time 721 ms). Volunteers' posture at the time of impact was found to be highly variable irrespective of the type of reactions. The exception was the frozen strategy for which volunteers raised their arms in a triangle to cover their face and head.

A total of 52 simulations (51 corresponding to the initial conditions from the experiments and one for the base condition corresponding to the standard walking posture) were performed using the MADYMOTM software package. For the initial conditions simulated (40 km / h impact), post impact kinematics and injury predicting parameters were found to be dependent on pedestrian pre-impact posture. In 10 of the simulated conditions, Head injury criteria (HIC) were found above the human head tolerance limit value typically used (i.e. HIC=1000), indicating chances of severe head injuries. As compared to the simulation for standard walking posture (HIC =1114), HIC was higher in 5 simulated conditions (with a maximum value of 1 327) whereas it was lower in the rest of the cases (i.e. in 46 cases). It implies that as far as head injuries are concerned, the standard walking posture commonly used for passive safety evaluation may not be the most critical / severe posture. However, it could represent a 90th percentile in severity, which may be acceptable for protection design. One of the possible ways to account for the most severe cases during the passive safety evaluation of a vehicle could be by setting a target HIC lower than the HIC estimated for the standard walking posture.