Machine learning System that can quickly detect traffic accidents and congestion with TensorFlow

Because Iowa state in the United States is an area where snow is accumulated about 33 cm on average annually, it is important to keep the road safe, especially in the winter. In Iowa state where car accidents increase rapidly in winter every year, in order to improve safety and efficiency in roads in winter, I analyze traffic situation using machine learning and identify traffic jams and occur accident quickly We developed a system that can detect.

When Iowa's snow piles up, TensorFlow can keep roads safe
https://www.blog.google/technology/ai/when-iowas-snow-piles-tensorflow-can-keep-roads-safe/

Keeping Iowa roads safe with AI - YouTube


"As many people say, the weather will change if you wait 10 minutes, even if you do not like the weather of Iowa," said Jeff Werrick.

In such Iowa state where the weather changes, Mr. Welic is working as a bus driver for the Roland Story Community School District.

Matt Patton, Chief Executive Officer of Rowland-Story Community School District, said, "Many children in Iowa traveled a distance of 10 to 12 miles (16 to 19 km) to attend school I am particularly concerned about school principals because children are safe schools for children. "

"When I get up at 4:40 in the morning, I usually go outside and check the state of the road. I and Matt are going to check the road and I'm in charge of the other side of the road that Matt is in charge of "Mr. Welic.

Mr. Patton said, "It is very important for us to keep in mind what is happening on the Iowa Road even if the weather changes," Iowa's student's long distance to school is safe In particular, I have revealed that I have done detailed checks on road conditions especially during winter.

Subsequently, Mr. Bonnie Castillo, who works as a director at the Traffic Management Center of the Iowa State Transportation Authority, appeared. "The goal for the Iowa State Bureau of Transportation is to deliver travelers to the destination efficiently and safely.We constantly check information coming from over 400 cameras and other data sources 24 hours a day, 365 days a day It is very difficult to analyze the wide variety of data we enter into, "he says about the work of the Bureau of Transportation.

A lot of monitors are lined up in the traffic management center inside the transportation office, and here I check the traffic situation of Iowa state while checking the information sent from the camera every day.

Tracey Bramble, Information Specialization at Iowa State Transportation Bureau, said, "We hope to summarize the traffic flow data flow with the help of the State of Iowa and be able to analyze more quickly "We are preaching the importance of the system for smarter analysis and use of the large amount of data dealt with by the Bureau of Transportation.

Therefore, what was done between the Iowa State Transport Authority and the Iowa State University's Traffic Research Institute is a project that uses " machine learning " to improve the safety and efficiency of winter roads.

"We are very excited about the idea of using machine learning for traffic management, Google's TensorFlow helps to understand traffic patterns and information," said Neil Hawkins, director at the Traffic Laboratory. And we will be able to issue warnings etc. from more accurate forecasts. "

Hawkins and his colleagues analyzed the vast amount of data dealt with by the Bureau of Transportation and developed a system to detect accidents and traffic jams as soon as possible.

Tingting Huang, a research scientist at Iowa State University who worked on development, said, "Using TensorFlow and machine learning, we automatically detected" whether a traffic accident occurred "from images collected from as many as 400 cameras We developed a system that can transmit information to transportation engineers in the event of a traffic accident. "

By analyzing visual data gathered from fixed cameras and cameras attached to snowplows, it is possible to identify traffic congestion and respond to accidents more quickly. Furthermore, Iowa state is not the only one trying to improve the traffic situation by combining machine learning and TensorFlow. In California where so much snowfall does not occur, TensorFlow was used to alleviate traffic congestion .

"If technology helps us judge what is happening throughout the school district, we should be able to make better decisions for children," Wellik said.

Mr. Patton said, "To be honest, if you do not need to drive in order to get up in the morning to check the road conditions, you will not get up before 5 o'clock in the morning," Patton said about the benefits of the system with a laugh I will.

If Mr. Hawkins said "machine learning has provided a revolutionary way of connecting our data with the cars ..." ...

Mr. Castillo said that "Iowa's roads are better thanks to the technology we provide for safer and more efficient traffic conditions" that we have, "said Machinery Learning and TensorFlow has said that we were able to greatly improve the traffic situation of Iowa in winter.

The Iowa State Transit Authority also states that machine learning can be used to better predict the occurrence of traffic accidents and the reasons for accidents. Also, during the winter season, traffic conditions are bad and commuting time tends to be long, but we encourage you to refrain from using cruise control even during the storm.