Why Your City’s ITMS is Underperforming?
Most congested Asian cities are Indian (TomTom 2025)
Most congested Asian cities are Indian (TomTom 2025)
Most congested Asian cities are Indian (TomTom 2025)
Here is an uncomfortable truth that no one in the traffic management industry wants to say out loud: most Intelligent Traffic Management Systems deployed across Indian cities are not managing traffic. They are watching it. And there is a world of difference between the two.
India spent billions of rupees deploying ITMS infrastructure under the Smart Cities Mission. Cameras went up at major junctions. Command centres were built. Dashboards were installed. But by the government’s own assessment, only 30 of 100 smart cities are actively using their systems for traffic management. The rest? The screens are on, but the intelligence is missing.
This article is not an attack on ITMS. It is a diagnosis. If you are a traffic commissioner, a municipal engineer, or a smart city project director, this piece will help you understand exactly why your system is underdelivering—and what the emerging alternative looks like.
The Camera Paradox: More Eyes, Less Insight
The foundational assumption behind most ITMS deployments is simple: if we install enough cameras at enough junctions, we will understand traffic. This assumption is wrong—not because cameras are bad technology, but because they are solving the wrong problem.
Consider what a typical ITMS camera can do. It can count vehicles passing a specific point. It can capture licence plates for enforcement. It can provide a live video feed to a control room. What it cannot do is tell you how fast traffic is actually moving on the 3 km stretch between that camera and the next one. It cannot tell you whether the congestion at Junction A was caused by an event at Junction B, two kilometres upstream. It cannot tell you that tomorrow at 5:47 pm, this exact road will have a 22- minute delay because it is the last working day before a long weekend.
This is not a theoretical gap. A typical Indian city’s ITMS covers the intersections where cameras are installed—roughly 15–20% of the road network. The remaining 80% is a blind spot. The traffic flowing through residential streets, arterial roads between junctions, and secondary corridors simply does not exist in the system’s view of the world.
Four Ways ITMS Is Failing (With Real Examples)
Let us move beyond generalities. Here are four concrete, documented failure
patterns that show up across ITMS deployments in Indian cities:
1. Optimised for fines, not for flow
In Bhopal, the ITMS was deployed as a flagship smart city project. Cameras went up at strategic junctions, a command centre was built, and the system began generating e-challans. But here is the telling detail: the entire investment was architected around catching violations—not around understanding or reducing congestion. No module was deployed to measure corridor travel times. No algorithm predicted where the next jam would form. The cameras could tell you that a motorist ran a red light, but they could not tell you why the 3 km stretch behind that junction was gridlocked every evening. The roads remained just as congested. The system was technically “working,” but it was solving for compliance, not for mobility.
2. Data that cannot be acted on
Most ITMS deployments generate enormous amounts of video data. But video data is not traffic intelligence. A control room operator watching 40 camera feeds cannot synthesize that into a decision like “increase green time by 15 seconds on the southbound approach for the next 3 cycles.” The data sits in silos—camera by camera, junction by junction—with no analytical layer connecting it into a network- wide picture. Officers end up making the same gut-feel decisions they made before the cameras went up.
3. No predictive capability
This is the most consequential gap. Every ITMS in the country is reactive. It shows you a traffic jam that has already formed. By the time a control room sees the red on screen, the queue is already 2 km long and 15 minutes old. There is no system in production today that tells a traffic officer: “In 20 minutes, MG Road will have severe congestion because a cricket match at the stadium is ending and 30,000 people will try to leave simultaneously.” Reactive systems cannot prevent congestion. They can only document it.
4. No before-and-after measurement
Perhaps the most damaging gap: ITMS cannot prove its own value. If a city widens a road or adjusts signal timings, there is no mechanism within the existing system to measure whether travel times actually improved. Without a baseline and continuous measurement, every infrastructure investment is a leap of faith. Commissioners cannot justify budgets with data, and engineers cannot prove their interventions worked. The system that was supposed to bring accountability to traffic management is itself unaccountable.
The Root Cause: ITMS Was Built for Enforcement, Not
Management
Here is the core insight that most vendors will not tell you: the ITMS architecture was never designed for traffic management. It was designed for traffic enforcement. The entire system—cameras, ANPR, e-challans, licence plate recognition—is optimized for catching violations, not for understanding traffic flow.
This is not a flaw; it is a design choice. Enforcement is valuable. Red-light violations, speeding, and helmet compliance genuinely matter for road safety. But somewhere along the way, cities began treating an enforcement system as a management system. They are fundamentally different problems requiring fundamentally different architectures.
| Feature | Capability Enforcement System (ITMS) | Management System |
|---|---|---|
| Coverage | Junctions with cameras (~20% of network) | Entire road network, every street |
| Data Type | Vehicle counts, licence plates, video | Travel times, speeds, congestion patterns |
| Update Frequency | Continuous video (manual analysis) | Every 2 minutes, automated analytics |
| Prediction | None — reactive only | 15–30 minute congestion forecasting |
| Actionable Output | Challans, live video feeds | Signal timing recommendations, bottleneck ranking, root cause analysis |
| Deployment Time | 2–5 years (hardware installation) | Weeks (software deployment) |
| Impact Measurement | Cannot measure before/after | Continuous baseline comparison |
This table is not theoretical. It describes the difference between what cities deployed and what they actually needed. The tragedy is not that ITMS was a bad investment—it is that it was an incomplete one.
What Would a Real Traffic Management System Look Like?
The table above tells you what is missing. But a true traffic management system does not just fill those gaps—it unlocks capabilities that were never on the table with lepton.ai/blog Lepton Maps · Page 5 cameras alone. Once you have network-wide, real-time speed data on every road, updated every two minutes, the possibilities go far beyond monitoring:
Simulate before you build. With months of historical travel-time data on every road, a city can simulate the impact of a proposed flyover, a lane closure, or a one- way conversion before a single rupee is spent. What happens to surrounding corridors if you close this road for metro construction? Which alternative routes will absorb the load, and will they cope? Today, these questions are answered with gut instinct and trial-and-error. With a data-rich intelligence layer, they become modelling exercises with measurable answers.
Plan public transport with real demand data. Origin-destination data reveals where people are actually travelling—not where planners assume they travel. This transforms public transit planning from a political exercise into an evidence-based one. Where should a new bus route run? Which metro feeder corridors are underserved? Where does the last-mile gap actually exist? These are answerable questions once you have continuous, city-wide travel flow data.
Prepare for events and disruptions, not just react to them. A cricket match at a stadium, monsoon flooding on a key arterial, a VIP motorcade during rush hour—all of these are predictable disruptions. An intelligent system fuses traffic data with event calendars, weather forecasts, and historical patterns to pre-position resources and adjust signal plans hours in advance. The shift is from crisis management to planned response.
Diagnose root causes, not just symptoms. A camera shows you a jam at an intersection. A management system tells you that the jam is caused by a bus stop 400 metres upstream where passenger boarding blocks the left lane during peak hours. Or that weekend congestion at a flyover approach correlates with parking overflow from a nearby commercial complex. The ability to correlate congestion patterns with their upstream causes is what turns a traffic department from firefighters into urban planners.
The Path Forward: Upgrade the Brain, Not the Eyes
The good news is that cities do not need to rip out their existing ITMS infrastructure. Cameras still serve an enforcement purpose. The investment is not wasted. What cities need is an intelligence layer that sits on top—a “brain” that ingests data from lepton.ai/blog Lepton Maps · Page 6 cameras, probe data, weather systems, event calendars, and GIS layers, and turns it into decisions.
Think of it as the difference between having a stethoscope and having a diagnostic system. The stethoscope (camera) gives you raw input. The diagnostic system (intelligence layer) tells you what is wrong, predicts what will go wrong next, and recommends a treatment.
This is already happening. In cities where software-first platforms have been deployed alongside existing ITMS, the results are tangible. Traffic police who previously spent hours manually tracking congestion on consumer Google Maps now receive automated, predictive alerts. Daily reports that took a team of people to compile each morning are now generated and delivered automatically to officers’ phones before they start their shift. Chronic bottlenecks that persisted for years are being diagnosed to their root causes for the first time—and the causes are often surprising: a bus stop placement, a parking overflow from a nearby mall, a signal phase that has not been updated in a decade.
What You Should Do Next
If you are responsible for traffic management in your city, here are three questions to ask your team this week:
- What percentage of our road network does our current system actually cover? If the answer is below 50%, you have a visibility problem, not a technology problem.
- Can our system predict a congestion event before it happens? If the answer is no, your system is documenting traffic problems, not solving them.
- Can we measure the before-and-after impact of a specific intervention? If the answer is no, you have no way to know whether your budget is being well spent.
The cities that will lead in traffic management over the next decade will not be the ones with the most cameras. They will be the ones with the best intelligence. The infrastructure race is over. The intelligence race has just begun.
Sources & References
TomTom Traffic Index 2025 — Global congestion rankings and hours lost per city.
India Economic Survey 2025–26 — Congestion pricing and urban mobility reform data.
Citizen Matters: Bhopal’s ITMS shows few signs of intelligence — System deployment priorities and effectiveness data.
Down to Earth: After a decade, only 18 of 100 smart cities completed — Smart Cities Mission status.
Swarajya Mag: Bengaluru’s Rs 20,000 crore traffic losses — Economic cost estimates.
Umang Saraf is the Director of Innovation Labs at Lepton Maps, where he leads TraffiCure—an AI-powered traffic intelligence platform built on Google’s Roads Management Insights. He works with traffic police and municipalities across India and the Middle East to shift cities from reactive traffic monitoring to predictive traffic management.
