American scientists can predict car accidents before they happen - ForumDaily
The article has been automatically translated into English by Google Translate from Russian and has not been edited.
Переклад цього матеріалу українською мовою з російської було автоматично здійснено сервісом Google Translate, без подальшого редагування тексту.
Bu məqalə Google Translate servisi vasitəsi ilə avtomatik olaraq rus dilindən azərbaycan dilinə tərcümə olunmuşdur. Bundan sonra mətn redaktə edilməmişdir.

American scientists can predict car accidents before they happen

Today's world is one big labyrinth, connected by layers of concrete and asphalt that allow us to travel by car. In many of our traffic-related advances, GPS allows us to use fewer neurons through mapping applications; cameras warn us of potentially dangerous holes and potholes, and electric autonomous cars have lower fuel costs. But accidents, alas, still happen. Scientists managed to develop a model for predicting the risk of road traffic accidents (RTA). The publication spoke about this in more detail. MIT News.

Photo: Shutterstock

To avoid the uncertainty associated with accidents, scientists at the Massachusetts Institute of Technology's Computer Science and Artificial Intelligence Laboratory (CSAIL) and the Qatar Center for Artificial Intelligence have developed a deep learning model that predicts the risk of accidents.

Based on a combination of historical crash data, road maps, satellite imagery and GPS footprints, risk maps describe the expected number of crashes over a period of time in the future to identify high-risk areas and predict future accidents.

Typically, these types of risk maps are recorded at a much lower resolution, ranging in the hundreds of meters, which means important details are not visible as roads become blurry. These maps, however, are grid cells of 5 × 5 meters, and the higher resolution gives new clarity: Scientists have found that, for example, highways have a higher risk than country roads, and highway exits are much more risky than others. sections of the road.

On the subject: Fatal crashes and huge traffic jams: worst US cities for motorists

“By learning the underlying risk distribution that determines the likelihood of future crashes in all locations, we can determine safe routes even without data from previous crashes. This could in the future allow auto insurance companies to provide customized insurance plans based on customers' driving trajectories. Or help planners design safer roads to avoid future accidents,” said MIT CSAIL graduate student Songtao He, one of the study’s authors.

Although car crashes are relatively rare, they cost about 3 percent of global GDP and are the leading cause of death for children and young people. Previous attempts to predict the risk of an accident were largely "historical", as the area would only be considered high risk if a previous accident had already occurred nearby.

The team's approach provides a wider network for collecting critical data. It identifies high-risk locations using GPS trajectory templates that provide information on density, speed and direction of travel, as well as satellite images describing road structures such as the number of lanes, the presence of a shoulder or a large number of pedestrians. Then, even if no accidents were recorded in a high-risk area, it can still be identified as a high-risk area based only on its traffic patterns and topology.

To evaluate the model, the scientists used crash data for 2017 and 2018 and tested its performance in predicting crashes in 2019 and 2020. Many locations were identified as high-risk areas even though they had no reported accidents in previous years.

You may be interested in: top New York news, stories of our immigrants and helpful tips about life in the Big Apple - read it all on ForumDaily New York

“Our model can be generalized from one city to another by combining multiple factors from seemingly unrelated data sources. This is a step towards general AI because our model can predict accident maps in uncharted territories,” says Amin Sadeghi, lead researcher at the Qatar Computing Research Institute (QCRI) and author of the paper. “The model can be used to produce a useful crash map even in the absence of historical crash data, which can be positively used for urban planning and policy making by comparing estimated scenarios.”

The dataset covered 7500 square kilometers from Los Angeles (California), New York (New York), Chicago (Illinois), and Boston (Massachusetts). Among the four cities, Los Angeles was the most unsafe, as it had the highest accident density. It is followed by New York, Chicago and Boston.

“If people have the ability to use a map to identify areas of the road that are potentially at risk, they can take proactive action. Apps like Waze and Apple Maps have incident management tools, but we try to get ahead of accidents before they happen,” he says.

He and Sadeghi co-wrote the article with Sanjay Chawla, Director of Research at QCRI, and MIT electrical and computer science professors Mohammad Alizadeh, Hari Balakrishnan, and Sam Madden.

Read also on ForumDaily:

What Americans are most afraid of

Five different Americas: what you need to prepare, going to the USA

How to live to be 120: advice from centenarians from around the world

Turn off the whole world with one button: what happens if the GPS stops working

On a rainy day: who are the preppers and what can they teach during a pandemic

model Educational program accidents
Subscribe to ForumDaily on Google News

Do you want more important and interesting news about life in the USA and immigration to America? — support us donate! Also subscribe to our page Facebook. Select the “Priority in display” option and read us first. Also, don't forget to subscribe to our РєР ° РЅР ° Р »РІ Telegram  and Instagram- there is a lot of interesting things there. And join thousands of readers ForumDaily New York — there you will find a lot of interesting and positive information about life in the metropolis. 



 
1070 requests in 1,081 seconds.