InDeV

Page 1

The road to improved traffic safety Many factors may be involved in a road traffic accident, including lighting, street layout or sight obstructions. We spoke to Aliaksei Laureshyn about the InDev project’s work in looking deeper into the underlying causes behind road accidents, research which could inform the development of effective counter-measures to improve safety A pedestrian or cyclist involved in a road traffic accident is often much more badly affected than the driver of the vehicle, who is relatively well protected by the surrounding metal framework. Deeper analysis of accidents could lead to new insights into the underlying causes behind them and so help to improve the safety of vulnerable road users (VRUs), a prime motivation behind the work of the InDev project. “The idea was to look deeper into the causative factors behind those accidents,” explains Aliaksei Laureshyn, the manager of the project. This includes not only analysis of accidents and how they develop, but also research into near misses and other, closely related areas. “We are also looking into traffic conflict studies and behavioural studies. Before an actual accident happens, there might have been several near-misses. Maybe the driver avoided a collision by half-a-second, but otherwise the circumstances around how the incident developed are often the same as with an actual accident,” says Laureshyn. This area of research has a long history, dating back around half a century, when observers would be sent out to watch the roads and gather more data about the root

a)

causes of traffic accidents. Technology has since moved on significantly of course, now Laureshyn and his colleagues in the project are applying modern techniques to help develop more rigorous, detailed methods of analysing the circumstances that lead to traffic accidents. “We are using data from video cameras, and also we are trying to develop computer vision tools that can help in detecting certain road traffic situations,” he outlines. With

Vulnerable road users This work is built on a recognition of the shortcomings of previous traffic conflict analysis methodologies, which were designed primarily with cars in mind, and didn’t always take vulnerable road users fully into account. “For example, if you imagine two cars travelling at 40 kph avoiding each other, you would say this is a serious conflict. But then if instead of a second car there is a cyclist, then intuitively you would say that it’s much more

Before an actual accident happens, there might have been several near-misses. Maybe the driver avoided a collision by half-asecond, but otherwise the circumstances around how the incident developed are the same as with an actual accident modern technology, researchers are able to gather large volumes of information on traffic over extended periods. “We collect accident data, conflict data, and more,” continues Laureshyn. “The project is about methodological research. We aim to build a deeper understanding of how accidents and conflicts should be analysed before we can recommend a methodology for wider use by local municipalities.”

dangerous, as the cyclist is less protected and so more vulnerable. We’re trying to include that aspect of vulnerability, so that estimates of the severity of a conflict are sensitive to who is actually involved,” outlines Laureshyn. Several sources of data are combined in the project, including accident databases, behavioural data and surrogate safety indicators, with researchers looking to develop a toolbox for traffic conflict analysis.

b)

b) special software detects the presence of the relevant road users and selects those with arriving close in time;

c) an expert investigates the near-misses using another tool and finally judges how close it was to a real collision

b)

a) Cameras are usually installed on existing infrastructure like lamp posts;

74

EU Research


“We’re developing the technical theory behind the toolbox,” says Laureshyn. There are several strands to this research, including the development of computer vision tools to monitor roads and extract relevant information. While a monitoring tool could be applied to gather data on a road over a relatively long period, researchers need to take the changing circumstances over that time into account. “Traffic levels change, levels of sunlight change and shadows move. Developing a tool that will provide stable levels of performance in those different conditions is a very challenging task,” acknowledges Laureshyn. This is a major challenge in the computer vision field, and while Laureshyn does not expect it to be completely resolved in the course of the project, he says some important advances have been achieved. “For example, we are developing simple tools that we can use to remove video at points where we know nothing is happening,” he says. “So if you put in 24 hours of video, it will reduce it to a more manageable amount.” A road traffic expert can then analyse the video excerpt, identify whether it can be classed as a near miss or a conflict, and gain insights into the underlying causes behind the specific incident. While there are of course aggressive drivers on the road, and some pedestrians and cyclists don’t obey the rules, ultimately nobody wants to put themselves at extreme risk. “When we see situations like that, it means there has been a breakdown – somebody has miscalculated,” says Laureshyn. There could be

a wide range of possible explanations behind an accident; Laureshyn says the project aims to put analysis on a firmer footing, giving professionals a basis on which to investigate road safety. “We are writing a handbook for practitioners, such as engineers and municipal authorities, where we describe different methods to study the safety issues facing road users,” he outlines. “It will be a step-by-step instruction on how to conduct a conflict study, including things like where to locate the cameras, and automated behaviour data collection.”

Preventative measures The wider aim in the project is to more closely link the factors that lead to accidents to the risks facing vulnerable road users. This will provide a more solid evidence base for the development and eventual implementation of preventative measures. “With our methodology, we aim to better understand how a road traffic incident develops. Is it because of light conditions or sight obstruction? Or are people driving too fast in that particular area? That kind of information is highly relevant when thinking about the implementation of countermeasures,” points out Laureshyn. The focus in the project has been on acquiring and analysing the relevant information, yet Laureshyn and his colleagues are well aware of the wider importance of their research, and are keen to have a practical impact on road safety. “We want the handbook, in a physical format, lying on the desktop of each person working in road safety,” he says.

b)

InDeV In-depth Understanding of Accident Causation for Vulnerable Road Users Project Objectives

The InDev project aims to examine how combining different methods can help researchers to build a deeper understanding of the underlying causes behind accidents involving cyclists and pedestrians. There is a particular focus in the project on methods that do not require accident data to assess the safety of a particular location. This could be behavioural or traffic conflict observations, the latter referring to a situation where an accident was avoided by only a small margin.

Project Funding

RIA - Research and Innovation action. Budget = EU contribution = 4.900.000 Euro Horizon 2020, Grant agreement No. 635895

Project Partners

Lund University, Sweden • Ålborg University, Denmark • The Federal Highway Research Institute, Germany • Hasselt University, Belgium • Netherlands Organisation for Applied Scientific Research, Netherlands • Warsaw University of Technology, Poland • Ingeniería de Tráfico SL, Spain • Institute of transport Economics, Norway • Polytechnique Montréal, Canada

Contact Details

Project Coordinator, Aliaksei Laureshyn Senior lecturer Lund University Sweden T: +46 462 229131 E: aliaksei.laureshyn@tft.lth.se W: www.indev-project.eu

Aliaksei Laureshyn

c)

www.euresearcher.com

c)

Aliaksei Laureshyn is a senior researcher at Lunds University in Sweden and Institute of Transport Economics (TØI) in Norway. His main research interest is the use of traffic conflicts and other surrogate measures to study safety in traffic, particularly for the unprotected road users (pedestrians, cyclist). He has been working closely with researchers within computer vision to develop tools for autoamted detection and classification of traffic conflicts.

75


Turn static files into dynamic content formats.

Create a flipbook
Issuu converts static files into: digital portfolios, online yearbooks, online catalogs, digital photo albums and more. Sign up and create your flipbook.