Software Divines the Risk of Road Accidents

Images: Universidad de Los Andes

A map showing the accident risk levels of Bogotá's streets

Cities are chaotic. The mix of foot traffic, cars, and buses means that some areas are hotbeds for traffic accidents. The city of Bogotá, Colombia is no exception, so Los Andes University computer scientists have come up with an application that translates accident statistics into dynamic maps of road accident risk on every street in the city. The maps can even identify areas where it'd be easiest to increase safety.

“The main conceptual issue was how to measure risk and the definition of work scales,” says Germán Bravo, a systems and computer engineer at the university  who presented the work at the IEEE Vehicular Technology Conference in Seoul. “We found that other groups work on risk, but not at the scales we wanted.” Most groups modeling accident risk work on either the micro-scale—individual intersections—or at the macro scale—districts or neighborhoods. But the Los Andes team wanted to delve deeper.

Their first step was to take accident statistics from Colombia’s Transportation Ministry and convert them to street locations, so the accident patterns could be grasped easily: “People don’t work in geographic coordinates,” says Bravo.

The researchers analyzed five large cities in Colombia, sorted the accidents by type, and then set out to determine the risk of individual areas. Their risk-determining algorithm had to combine knowledge of high-activity locations in the given cities, accident mitigators such as traffic lights or walkways, socio-economic distribution, and a matrix describing people’s general movement from place to place. The maps they came up with were broad, covering all five cities, but also specific enough to zoom down to specific streets and landmarks.

A "people attraction" map of Bogotà shows areas of high traffic.

One surprise along the way from the analysis were the “attraction maps.” These showed the power of different areas to attract traffic headed to or from places like bars, schools, malls, or businesses. The maps were a necessary step towards calculating risk, but they were intended as just an intermediate.

“At first the team did not believe in their utility, but later they loved them,” says Bravo. “When people see the maps they really recognize the city where they live—and they really help to explain certain accident cases,” such as the high incidence of accidents around schools in high-traffic zones.

Knowing the creators and reducers of risk plus the actual accident statistics, the application can generate maps showing accident hotspots and the locations with the most danger. It can even create maps showing the potential to improve risk in neighborhoods, pointing out places with the densest risk that might be amenable to changes in infrastructure.

Bravo says that the maps are already being used to identify areas where small investments of resources could dramatically reduce risk.

Right now the maps are used in partnership with a road-safety agency in Colombia, but Bravo hopes to work with the Transportation Ministry and also create a cloud or web accessible format to open the maps to everyone. With the help of these risk maps, he says, Colombia will be able to convert nebulous data to concrete conclusions about when and how accidents happen—and how to stop them.

Advertisement

Cars That Think

IEEE Spectrum’s blog about the sensors, software, and systems that are making cars smarter, more entertaining, and ultimately, autonomous.
Contact us:  p.ross@ieee.org

Senior Editor
Philip E. Ross
New York City
Assistant Editor
Willie D. Jones
New York City
 
Senior Writer
Evan Ackerman
Berkeley, Calif.
Contributor
Lucas Laursen
Madrid
 

Newsletter Sign Up

Sign up for the Cars That Think newsletter and get biweekly updates, all delivered directly to your inbox.

Advertisement