Taking account of how individuals react to a risk

Science topics February 2015 Road safetyHuman behaviourModelling and computer simulation

Thomas Robert, Researcher - TS2 Department , LMBC Laboratory

In most cases, a person who is facing an imminent risk (loss of balance, forthcoming impact, etc.) has the time to perceive it and react.
When a road traffic accident happens, drivers generally see the approaching danger. This gives them time to try to brake in order to avoid the accident or at least to stiffen up in order to better withstand the impact. In the same way, passengers who are standing in the Metro can lose their balance during emergency braking. But in most cases they will have time to try to regain their balance or at least try to limit the impact of a fall.
When these reactions are appropriate, they make it possible to avoid the accident, or limit its consequences. However, they can also have the opposite effect and generate an additional risk. A good example of this is when a pedestrian avoidance system1 is activated in an emergency and causes a secondary accident when the first could have been avoided otherwise.
Studying these reactions and integrating them within digital human models is therefore of major importance. The LBMC is therefore engaged in several research projects to better identify hazardous situations and better protect individuals.



Analysis of movement in pre-crash situations

How do pedestrians – young or elderly – react when a high speed vehicle suddenly appears when they are crossing the road and threatens to hit them? And what are the possible consequences of this reaction?
To answer this question, IFSTTAR has set up the following experiment :
Young and elderly subjects take turns to move around a virtual environment which simulates a street with traffic consisting of a flow of vehicles. When one of the pedestrians is crossing the street, the simulator generates a high speed vehicle, which is accompanied by the noise of a crash. The reaction of the pedestrians is recorded with a motion analysis device similar to those used in the video games or cinema animation industry.
This experiment collects data about the posture of subjects and their muscular activity2at the time of impact. This data is then integrated within the various injury prediction models in order to estimate whether these reactions tend to reduce or increase the risk to which the pedestrian is exposed.

Example of different postures observed at the time of the virtual impact.


Modelling balance control mechanisms


Keeping balance when standing is a relatively difficult task. This is particularly true in an unstable environment such as a public transport vehicle.
Falls are, in fact, one of the most common causes of injury. Elderly people often mention that the fear of falling prevents them from using public transport. This observation raises many questions: to what risk of falling is an individual exposed according to their capacities and the interference with balance they experience? Can they recover their balance, and if so how?
To answer these questions, the LBMC has been working for several years on modelling balance control and recovery mechanisms. The models it has developed consist of a mechanical representation of the human body and controllers which determine the necessary recovery actions3. They are then tested on previously collected experimental data collected in the course of experiments conducted on volunteers.



1. Example of the system developed by Toyota on http://newsroom.toyota.eu/pdf.do?id=3654



Find out more ...

2 Anurag Soni, Thomas Robert, Frédéric Rongieras, and Philippe Beillas. Observations on pedestrian pre-crash reactions during simulated accidents. Stapp car crash journal, 57:157–183, 2013.
3 Zohaib Aftab, Thomas Robert, and Pierre-Brice Wieber. Predicting multiple step placements for human balance recovery tasks. Journal of Biomechanics, 45:2804 – 2809, 2012.