Traffic signals have been in use ever since automobiles became the preferred mode of travel. Yet, little has been done to improve their functioning, in comparison with innovations in automobile design. A group at IIT Madras has studied Indian traffic conditions and come up with three counter-intuitive strategies to improve efficiency of oversaturated traffic flow through road intersections.
In a preliminary study, Radhakrishnan, at present an assistant transport planner with Atkins, and Gitakrishnan Ramadurai from the Department of Civil Engineering, IIT Madras, collected data on the variation of headway, which is the distance between the leader vehicle and the following one. Observing vehicles at the traffic signals near three places in Chennai (Tidel Park, Tiruvanmiyur junction and Adyar depot), they found that the distance between two successive vehicles (headway) increases with duration of the green signal. This implies that longer the duration of green signal, there would be less throughput, thereby reducing the efficiency of the signal. Developing on this study, Dr Ramadurai arrived at the first strategy: keeping the duration of green signals short enough that the headway reaches a saturation value, in his paper in Communication Systems and Networks, conference proceedings published by IEEE.
Bottleneck metering
The second strategy proposed in the paper is to have bottleneck metering along the lines of traffic control on expressways, near entry points and exits. Like highways, expressways have points where traffic leaves or enters through lanes, such as in T-junctions. These often experience bottlenecks because vehicles, not being in specific lanes, merge from different directions, thereby causing ‘turbulence’ in the traffic flow. The researchers collected data from arterial roads in Chennai where a four-lane one-way section of a road merges with a two-way section with two lanes in each direction in a T-junction. They monitored average flow of different types of vehicles before, during and after congestion for three days. They found that for two-wheelers and cars, there was an increase in flow once congestion starts, whereas for three-wheelers and heavy vehicles, the flow decreases. Overall, they found a capacity gain at the bottleneck, instead of a capacity drop expected from reported literature, and flows increase once congestion sets it. Taking such factors into account, they suggest that by placing signals separately for each lane that enters or leaves a bottleneck on an arterial road, and using green signals alternately, a higher throughput or a capacity gain can be engineered.
Storage space
The third strategy is based on the observation that two-wheelers have significantly lower headways compared with others; they also have negligible start-up delay. Also, in congestion, two-wheelers filter through the gaps and increase overall throughput. “We show that two-wheeler throughput can be increased significantly by providing them space upfront at a signalized intersection without affecting other vehicles significantly,” says Dr Ramadurai. The idea is to have separate lanes for two-wheelers close to intersections and an exclusive two-wheeler storage space at the front where they can collect while waiting for the signal to change. A microsimulation model calibrated by S. Siddharth and Dr Ramadurai which was published in Proceedings of the Urban Mobility India 2013, conference, is used to test this idea. The simulation model showed a reduction in delay time of about 50% for two-wheelers and an overall delay reduction of 20%.
“The complex interactions [between different types of vehicles] are broken down by physically separating two-wheelers from three-wheelers, cars, and heavy vehicles and this may be a reason for smoother flow resulting in reduction in overall delays,” explains Dr Ramadurai. However, he adds that real-world trials are required to validate this hypothesis since simulation models are only an abstraction of reality.