Artificial intelligence-assisted traffic signals soon

Artificial intelligence-assisted traffic signals soon

Traffic management in Bengaluru will enter the artificial intelligence era: cameras and computers network to study vehicle density and tell signals when to change

Experts say that changes in the volume of traffic must be observed and the signals be upgraded every six months. (Above) At Seshadri Road.

All 387 traffic signals in Bengaluru will soon use artificial intelligence and regulate traffic more efficiently, according to Additional Commissioner of Police (Traffic) B R Ravikanthe Gowda. 
Under the new system, cameras study the traffic density and decide on how much time to allow vehicles to clear a signal, he told Metrolife. 

As of now, only 35, or a little over 10 per cent of all signals are adaptive. “The rest have a fixed time cycle,” he says.  The new signals have in-built artificial intelligence units that calculate, assess and network with other signals automatically.

 “The cameras calculate vehicle density and communicate the details to a central processing unit which, in turn, decides how much time should be given at each junction,” says Gowda. 
The countdown timer has two purposes — it helps save fuel as the motorists can turn off their vehicles during the waiting time, and secondly, it gives them an idea about when they can move. The new system will make the process more effective, he says.

 Traffic expert M N Sreehari says the police must regularly reset the signal at each junction.
 “There is something called co-ordination of signals and they must be in sync. For instance, when the first signal shows green, the next junction should also switch to green so that vehicles move non-stop,” he says. 

 All signals have a ‘compiler box’ with 64 built-in functions. “The box can provide options, such as a free left turn. The traffic police have to study each junction and accordingly make modifications,” he says. 

Dr Ashish Verma, associate professor, transportation and systems engineering, Indian Institute of Science, says the traffic police lack technical knowledge, and may not have the ability to determine how much time a signal needs. 

 Changes in the volume of traffic must be observed and the signals upgraded every six months, he says. 

 “Changes in flow and pattern take place when police implement a one-way or landscape changes. These changes must be taken into consideration when setting the timing of signals. I don’t think this is being done at the moment,” says Verma.