AI to ease manned traffic management by Kerala police with real-time adjustment of green, yellow, red lights

Traffic signal lights in busy areas may soon get an artificial intelligence upgrade, with timings for green, yellow and red lights set to change dynamically based on real-time vehicle queues.

Keltron has proposed a new ‘Count and Classification’ system to the Transport Department that would automatically adjust signal durations according to congestion levels. The technology would eliminate the need for police officers to manually control traffic during peak hours, a task still common at busy junctions.

The system builds on data collected from AI cameras already installed across the state. Using visuals and vehicle counts sent from cameras at each junction to a central traffic control room, the software can determine which road is busier and extend the green-light phase accordingly. The estimated cost for implementing the system at one junction is around ₹5 lakh.

Trial runs of the technology have been completed successfully, prompting Keltron to move ahead with formal proposals. Officials say the system will help ease bottlenecks by allowing traffic on high-volume roads to clear more quickly.

The ‘Count and Classification’ setup is also a major part of Keltron’s ₹170-crore traffic management project in Nagpur, covering 174 junctions.

Alongside this, Keltron has suggested the introduction of an ‘anti-glare filter’ system to tackle a problem that arose after AI cameras were first installed — sunlight interfering with the ability to clearly capture drivers’ faces. The anti-glare solution, now under consideration by the police chief, is expected to improve the accuracy and efficiency of AI-based monitoring.