Green Light to artificial intelligence

With population and city growth, congested traffic has become more and more frequent. The flow of transport leads to congestion, delays, accidents and a strong environmental impact.

According to the National Population and Household Census 2010 (Argentina), the “conglomerate” called Gran San Juan has more than 471 thousand inhabitants, what makes it one of the main urban centres of the country. The demographical growth led to an increase in the local number of vehicles, thus producing more frequent traffic jam. In order to deal with this problem, a research team at the Instituto de Automática (INAUT, CONICET-UNSJ) designed a traffic light with sensors that regulate traffic automatically. The device has already been installed in the district of Rivadavia in San Juan, in an intersection of four complex roads.

During the rush hours in Gran San Juan, there is traffic jam from east to west or vice versa. Some arteries are really congested whereas others are almost empty. With an ordinary traffic light, the queue of traffic is so long in peak times that it takes more than two minutes to cross, while in the other arteries there are no vehicles.

“We put sensors bellow the road surface to measure vehicular traffic flow. So in the artery that is more congested, the green light time extends up to 40 seconds and in those ones without vehicles, it lasts five seconds only. Consequently, traffic flows automatically because this device works in real time”, Antonio Cunsulo, CONICET principal professional at the Instituto de Automática (INAUT, CONICET-UNSJ), comments.

For his part, Daniel Patiño, researcher at the INAUT, states that the problem was posed by the municipal government in order to solve it through “automatism”. The researchers studied the problem and realized that at a national and international level, the available systems are centralized, that is to say that there’s an office with a control unit responsible for directing traffic. This is useful for a large city but they are expensive systems. So in this case, the scientists decided to design something independent and autonomous that centralizes automatically the traffic in the corner of Sargento Cabral and Santa María de Oro in the district of Rivadavia.

The researchers first carried out a pilot experiment in the district of Rawson for three years and after that the district of Rivadavia asked for one device. “The results show that the average wait time in that traffic light lasts no more than 20 seconds, therefore favoring bus drivers and commuters”, the engineer comments.

The scientists explain that the traffic light in San Juan works as if it had an operator indicating: ‘this artery is so congested that it is necessary to extend green light time’ or as if there was a policeman directing traffic. The idea was to replace the person with sensors. If they detect that the vehicle flow increases in one artery, the green light time is extended. When the vehicle flow finishes, the traffic light immediately goes red.

“If wait time is reduced, commuters’ attitude change. After work, people normally want to reach their homes as soon as possible. Inefficient systems that make people wait and queue to get the red light when there are no vehicles, make people lose their temper. This system avoids this and saves fuel. A vehicle consumes more fuel when its engine is stopped than when it runs and brakes, therefore reducing environmental impact. There is technology and knowledge, it is necessary to have the interfaces to connect the productive sector with the research laboratories in which that knowledge is generated”, Patiño concludes.