What it takes to get Bengaluru moving

Can tech and AI save us from mad traffic? DH journalist Roshan Thyagarajan visits the police nerve centre working tirelessly to untangle the city’s vehicular snarls

Bengaluru: I ran over a traffic policeman. This was circa 2008. In my defence, the man who would eventually end up in hospital engaged in the jump-scare routine Bengaluru cops are notorious for. It didn’t help either that another policeman pulled at my shirt while I was astride my scooter. I lost control. My body went with the angle of the pull. My bike went the other way. Bang!

I was 21 or so at the time, and had just joined this newspaper. I was
riding to a friend’s place with my helmet on, not drunk, in a straight line, with lights on and all documents in place. But the pointless tug followed by the crash put me behind bars before the two men convinced their seniors to let me go. I only had to write a letter saying ‘I won’t run over a cop again’. 

The staff talks to junction jockeys.

I was let out, but my conscience had to live with the sound of that man’s scream, the strange softness of his body under my wheel, the feeling of the lathi against the back of my thigh and lower back. Worst of all, I had come to terms with a senior policeman saying that he would book me for other crimes and let me rot in prison. His nonchalant delivery made me feel uncertain about what happens in these places because of these men.

It is ironic then that I am walking into Bengaluru’s Traffic Management Centre (TMC) on Infantry Road to do a story on those I don’t trust.

I am shown a reality that doesn’t corroborate my imagination. People here are nice to a fault. Maybe it has something to do with the label (journalist) I walk with, but their niceness doesn’t seem disingenuous. 

It’s me. I am pessimistic. All the more when it comes to greeting the men and women in charge of fixing this city’s greatest dilemma: traffic.
I arrive late because of it.

Calm and chaos

The elevator at the Centre is replete with signs and stickers that belong in the ’90s. I am ushered to a wooden door with a brass plate saying ‘Mission Control Room’. Head police inspector Anil Purohit throws the door open with practised pizzazz. It’s a Friday morning.

I am expecting chairs to be hurled and people screaming across the floor. Nope, none of that here. All of the white-shirted-khaki-trousered men and women wear nonchalant looks. They are busy, but don’t look stressed. They register my presence, but I am immersed in the wall of screens by now. A window into Armageddon.

Even as Purohit’s genial disposition attempts to accommodate my awkward expressions of curiosity and dumbfoundedness, the screens light up with images from various parts of Bengaluru, with the locations at the bottom.

Drones are offering a view. Google Maps is a pipeline of deep red lines suggesting heavy traffic. Several WhatsApp groups are feeding the screen information. The in-house software is putting out congestion alerts. Real-time information from samaritans is popping up. Messages from inspectors, constables and on-field jockeys come thick and fast. ‘Severe traffic on…’, ‘vehicle breakdown near... ’, ‘public gathering on... ’ are some of them.

They are all neatly categorised, but I see no messages about tech park traffic and school traffic, obvious culprits when it comes to our daily snarls. Roads affected by such vehicular movement are attended to every day. These message boards are mostly for assimilating anomalies.

There are desks with three tabletop screens on them, and there are hundreds of desks. Then there are control rooms, and more walkie-talkie jingles, tech guys staring at codes, and a lot of action but little noise.

Sixty-six people work on the floor the size of a basketball court across three shifts. There are five floors in all, and each serves a different purpose.

City overview

For ease of monitoring, Bengaluru is split into four zones (east, west, north and south) across nine major corridors. These divisions are further split into 10 subdivisions, and the 50 police stations in existence handle their workings.

The department sends out approx 6,000 challans a day.

I am not typically anxious in the face of overstimulation. I am today. Perhaps it is because those around me work, walk and talk with an irrational degree of poise. They aren’t even remotely transfixed by the screens.

Even a short ride in Bengaluru seems like a miracle on most days, and getting an aerial view makes you ask how any of us get anywhere without grievous injuries. The sixth-most congested city in the world is everything it is made out to be, and worse. You feel the gravity of the problem and even get a sense of the Rs 20,000-crore loss incurred by businesses as a result of the choke, one which turns our average travelling speed to 18 km an hour.

...“Bengaluru’s road network is approximately 14,000 km, and they say it can support 50-60 lakh vehicles. As of today, we have approximately 1.3 crore vehicles, and 2,000-2,500 vehicles are added to this number every day…” Purohit is still talking because his knowledge of traffic in the city is endless. But when he gets to explaining new technology, he realises he is talking to a tech philistine. 

Tech talk

Purohit, the man with a PhD in philosophy, dumbs it down just enough before bringing on a member of the ASTraM (Actionable Intelligence for Sustainable Traffic Management) team to educate me.

A young lad strips ASTraM down with real-time examples. 

One of which — a brief saying there’s traffic congestion right outside the TMC — prompts an inspector to call out for validation of the problem. The inspector, who is told the brief is a test, isn’t amused. 

This segment of the ASTraM is built for constables and other traffic technicians on the streets to alert the traffic control room. Once they feed the information to their handheld device, the control room is informed, and inspectors in charge check if the information is valid, and inform the next in the chain of command, using mics in the room.

The course of action is decided on the floor of the TMC and the decision is relayed back to the men and women on site. This could be anything from creating a one-way to diverting traffic through other roads.

The young man moves on to other facets of the technology, which run entirely on artificial intelligence and machine learning. ASTraM is still in its infancy, having been inaugurated on January 14 this year. But it has learnt how to walk already — congestion alerts, dashboard analytics, special event management, incident reporting using an in-house mobile application and a whole lot more than my rudimentary brain could grasp.

The last bit I hear with clarity is that the ASTraM doubles up as a predictive tool. Essentially, the TMC can preempt a tricky situation and preempt protocols to alleviate stress on the roads.

Crowdsourcing data 

I check back in when it is revealed that Swiggy, Rapido, Zomato, Redbus, BMTC and other mobility services offer data to them, besides individual volunteers.

By the time my brain conjures up an argument against the ethics of disbursing data freely, the topic of adaptive signals comes up. They do mention that all the data which comes to them is encrypted and pushed into confidential servers.

Anyway, there are 35 adaptive signals in operation at major junctions already, and they self-regulate the time stamp on green lights. The software also allows the signal to detect the queue length at the
junction using infrared sensors, and should the jam extend beyond 750 metres, junction jockeys are alerted to go to the location and take stock. The information is relayed back to the control room with photographic evidence and a brief. In the interim, the control room identifies alternative routes, if any, to redirect the traffic. 

Speaking of which, they will eventually also have the e-ambulance system where registered ambulances are assisted with coordinated movement across the city. Meaning, ambulances will be in conversation with the control room for the duration of their trip, and if they happen to be stuck for more than 120 seconds, personnel will be deployed to trigger movement along these dense stretches.

Drones come in handy too, but their battery life is fairly short (two hours) and they need to have trained hands for usage. While there are other pain points to this project, it is offering a perspective to the department, which in the past rarely looked beyond archaic underground sensors and intuition. 

Now, with the arrival of AI, the hypothetical horizon wears a silver hue.

Systems v/s humans

Purohit has deployed an army of people to show me around and explain every component installed in the TMC to get Bengaluru moving. I am even shown the room where the traffic violation challans are printed. 

What precedes the yellow-paper room of nightmares is me watching AI pick out traffic violations happening at the Intelligent Traffic Management System (ITMS) room. A livestream from a camera comes on, the software then picks out the violation with a box on the now-frozen image, the image is captured, the number plate noted, and on comes the next violation, and on, and on. 

“There are close to 35,000-40,000 violations in Bengaluru every day, and we can only send out challans to 6,000-7,000 of them because we
(six on a shift) have to manually validate these violations,” says a
member of the team.  

“The system in place is good but it’s not foolproof because of the lack of standardised number plates, the glare of the sun, shadow, and so on. There are plenty of reasons why the system could make a mistake. That’s why we intervene.”

When I ask if this is a slideshow of violations from the previous day, one of the system handlers laughs and says, “No, sir, these are real-time violations.” 

Brrr, the printer’s metronome breaks the spell of my wandering thoughts — my shenanigans of the past largely. 

My exposed feet are beginning to feel a bit cold from sweat as I smile and wave at the ITMS team. They smile back awkwardly. Once I exit
the sanitised room, I put on my shoes and notice a let’s-get-out-of-here spring in my step.

Meeting the chief

I am then ushered into a large room by a mesh of hands. It is a room full of trophies, glass tables, papers, framed certificates of yore... The Joint Commissioner of Police (Traffic) stands at his desk, looking as busy as I expected.

Sitting across from M N Anucheth, I experience something I haven’t before: reassurance, comfort, hope. I sense his desire to alleviate the biggest bane of the city. He is vulnerable enough to admit the department’s failings just as he is proud to delve into the measures they are taking. Besides transparency, he offers pointed information to my queries — “What’s the future? What’s the solution? When will it be fixed? What part does poor infrastructure play in all this? Is AI everything that you’re making it out to be?”

He is patient, kind even, in answering. Anucheth is building his legacy on the shoulders of those who preceded him. For that, though, he studies the city as he would his own palm lines to decipher what is happening.

He has some answers for now. What I hear isn’t resignation; it is hope, a thinking man handling an irrational predicament.

But, let’s face it, ‘you are not stuck in traffic, you are the traffic’. This 
Cartesian sentiment insists that we as the public should realise our folly too.

My visit ends and I am on the road, riding my bike. I am stuck at a traffic light near the Golden Palm Army Canteen for 60 seconds around 3 pm. I am cut off by someone near Anil Kumble Circle. I watch pedestrians dart across the zebra crossing at the Brigade Road junction before their turn comes. I sneer at people walking right down the middle of Church Street. I reach my office on M G Road.

All it takes is 20-odd minutes of Bengaluru to relegate the work the department is doing to a distant memory. I am livid with myself. I hope you can find it in you to be better than me. 

Like this story? Email: [e-mail]