Bengaluru's Traffic Choke | US Thanksgiving, Elephants Crossing Road Among Causes: JCP on Data, AI Findings

Bengaluru traffic, which hit international headlines as Trevor Noah’s show was cancelled last year, is now getting tech and data-driven solutions.

The Bengaluru Traffic Police have started to collect real-time data from 43 tech parks, map providers, aggregators and drones to intervene quickly and build a model that can predict traffic gridlock, especially in the IT corridor. The police are using artificial intelligence (AI) and machine learning to analyse this data and build a robust database.

However, the data crunching has also led to some strange reasons behind the traffic flow. For instance, during the Thanksgiving weekend in the United States last year, the city saw peak-hour traffic shifting from the usual time — instead of the 8-11am slot to after 11am — although it was during middle of the week. The traffic police believe this could be because the IT crowd, which had a holiday, stepped out around noon for other activities. “We have a planned calendar of events for India, but we realised we need to incorporate the US or other regions, as Bengaluru services those places,” said MN Anucheth, Joint Commissioner of Police (JCP), Traffic.

Recently, the NICE road saw traffic congestion around 5:15AM. On checking further, the police realised that a herd of elephants was crossing the road. “We want to understand how traffic flows on a volumetric level. We have taken data from various sources. All this is ingested by AI, which gives alerts. We are getting an alert every 15 minutes, so we can intervene accordingly. We are also looking to develop a baseline model, as to which road can handle how much traffic. Once we get good amount of data with learning, we will be able to go for a predictive model in four-five months,” said Anucheth.

The police are also getting granular details such as the number of cab rides being booked towards the tech park and data on vehicles parked in tech parks by 2pm, to help them assess the traffic likely to hit the tech corridor during evening peak hours. “We are getting data from aggregators on the kind of bookings to the tech park, or how vehicles will move from say Silk board to Marthahalli. From all major tech parks, we are able to understand the numbers of cars in the parking lots at 2pm. We are able to understand the exit from these tech parks which will contribute to 70% of the traffic load,” he said.