The Electronics and Telecommunications Research Institute (ETRI), a state research body, said in a statement on Tuesday that SALT analyzes real-time traffic data using cloud computing servers to speed up the simulation process.
Previous machine learning-based traffic congestion prediction technologies require researchers to create road simulation models each time. SALT is about 18 times faster than SUMO (simulation of urban mobility), a technology widely used to predict traffic.
"The cost of traffic congestion in South Korea is estimated at about 30 trillion won each year and it is on the rise. I hope this technology will help lower the socioeconomic cost of traffic congestion," ETRI research team head Min Ok-gee was quoted as saying.
SALT evenly dissects a target area to measure traffic. Researchers tested SALT to analyze and predict traffic by dissecting a district in southeastern Seoul, into 13,000 sections to simulate a 24-hour-long traffic flow in just five minutes.
The research team cooperated with Seoul City to come up with the most effective traffic light signal sequence. When actual changes were made to the traffic light sequence, the speed of traffic increased by 4.3 percent. SALT can be utilized in the crackdown of illegally parked vehicles and the detection of chronically congested road sections.