How this three-year-old Bengaluru startup is helping make US roads safer with its video analytics solutions


From the bylanes of Bengaluru to the highways of the United States, one startup is making global inroads into the world of video analytics-based road safety solutions.

The story of LightMetrics began in in February 2015 with six tech experts who worked together at Nokia Research for five years. Between them, they had filed over 90 patents and helped ship out millions of phones. They wanted to use their combined expertise in computer vision, machine learning, camera technology and signal processing to solve real challenges on the ground. That’s how Soumik Ukil, Ravi Shenoy, Mithun Uliyar, Gururaj Putraya, Pushkar Putwardhan and Krishna A G came together to start LightMetrics.

A video analytics company, LightMetrics helps transportation fleets, auto insurance companies and ride-sharing players better understand driving behaviour to improve safety and increase efficiency. The Bengaluru-based startup has a subsidiary in San Jose, California.

Making driving safer and efficient with video analytics

“Being first-time entrepreneurs, everything was a challenge, and anything that we thought of as easy actually ended up being a challenge,” says Krishna A G, Co-founder and Head of Business Development (North America) at LightMetrics. And like many first-time tech entrepreneurs, their key challenges primarily related to customers – their needs and access. “How do we ensure that the problem we are solving is relevant for the market? Given that the market abroad is more receptive to our products and solutions, how do we access customers there? These were the kind of questions we pondered over at boardroom meetings and even during coffee breaks,” says Krishna.

Three years into their journey, they have two key offerings aimed at the self-driving rental market and trucking companies. While RideView, a software development kit (SDK) offers real-time, on-device visual analytics which transforms a smartphone into a smart dash cam, RideCam is an affordable dashcam that leverages the computing power of a mobile device.

RideView has Android, iOS, and custom SDKs with offerings aimed at management of driving behaviour for fleet management, auto insurance, and for camera OEMs to make existing cameras smarter. The solutions helps measure key performance indicators for drivers in terms of safety and compliance, thereby enabling the identification repeat offenders so that they can be coached properly, as well as accident mitigation through real-time ADAS warnings. It also helps with real-time crowd-sourced mapping of roads, helps insurance underwriters understand risk better and reduce risk through positive driver behaviour modification, and gives Camera OEMs newer recurring revenue channels by enabling them to add services related to driver and vehicle safety.

On the other hand, RideCam is a cross-platform connected dashcam solution. It pairs automatically with mobile devices in the vicinity and streams 720p video in real time. The devices are powered by RideView SDK for ADAS notifications and driving behaviour analytics including event videos.

How customer feedback defined the product development journey

While it was RideView that was LightMetrics’s first product which was released, the market reaction and feedback led them to launch RideCam.

Krishna says, “Until two years ago, the Indian market wasn’t ready for a product like RideView. So we knew we had to focus on either the European or the US market. And we chose to focus on the US, considering it had a slight edge as a tech-friendly market. During our conversations with prospective clients we realised that drivers of large and heavy-duty vehicles don’t prefer mounting their smartphones to be used as a dashcam. They wanted a robust device that could be mounted independently. So that learning was instrumental in us developing our second product RideCam. What we essentially did was that we still leveraged computing power of an available device like a smartphone or a tab, but moved the camera out.”

This was a year ago. “We were running extensive beta pilots. Our potential and current partners have been using early prototypes for several months now. With the FCC (Federal Communications Commission) certification done, RideCam is commercially available in the US starting from February 2018.” He adds, “Fleet owners have already invested in tablets and smartphones as a productivity tool for drivers and to comply with federal regulations (ELD mandate). By leveraging the computer and modem on these mobile devices, RideCam is able to provide a terrific return on investment to its end-users.”

Team LightMetrics

Tapping opportunity with the right focus

According to Krishna, currently the team’s biggest focus area is driving business in the US market. “We will be strengthening our engagements with current and prospective customers and partners in the US. Interestingly, LightMetrics is seeing a lot of in-bound requests from Europe. So, we will be catering to these as well.”

Being based in India, do they have anything at all for the Indian market? The team says that they will shortly, as they are positive about the prospect that the current Indian market holds and are working on the sidelines to develop a product to suit local needs.

Winning support and validation

Being a software company, LightMetrics doesn’t manufacture hardware for RideCam. They work with manufacturing partners. Krishna says, “Often, manufacturers or large enterprises who prefer manufacturing the hardware through their OEM partners often ask for a camera reference design. So, we wanted to develop a reference design derived on our learnings so far.”

At this point, they applied for Cycle II of Qualcomm Design in India - Challenge 2017 (QDI-C 2017) and were among the six companies shortlisted. LightMetrics is among the six startups receiving mentoring and technology support as part of the run-up to the finale. Explaining how Qualcomm has become an active enabler in their startup journey, Krishna says, “We had the concept for the camera reference design ready and were looking forward to build the actual hardware product – the reference camera. QDI-C 2017 happened at just the right time for us.”

The engagement with Qualcomm not only enabled LightMetrics to build the camera reference design, but also enabled them to benchmark their software solutions on a number of Qualcomm platforms and also understand how they could enhance performance, speed, stability, power consumption by integrating Qualcomm technology.

Krishna says, “On the hardware front, Qualcomm validated the camera reference design for us, which is a huge deal given their massive experience with hardware and designing such systems. On the software side, they provided us with detailed inputs to optimise our algorithms for a particular chipset. We are still working with them on this front, trying to ensure that we get the maximum performance from their hardware.”

They have finalised the Qualcomm® Snapdragon™ 820 Mobile Platform for integration with the camera reference model. He says, “We want to be future-ready. In the time to come, we plan to add more features, more camera support, we want our hardware to be able to support higher computing with less power consumption, and that is why a high-power processor like Snapdragon 820 is important for us.”

Krishna says it is not the first time they have chosen to go with Qualcomm platforms over others. “Early on in our development, subconsciously we found ourselves recommending Snapdragon-based phones to our prospective customers for RideView and RideCam. The explanation basically came down to reliability, quality and having devices at different price points with a good user experience. And, LightMetrics too prides itself on making video telematics available at different price points, with different hardware options. So, it was a natural synergy.”

Automobiles and an AI-driven future

Talking about the automobile industry where connected vehicles are said to be the future, Krishna says, “Today the humungous amount of data that gets generated is most often used for predictive diagnostics, enhancing driver and road safety, performance, or driving efficiencies in a number of use-cases like usage of fuel, etc. But this is barely touching the surface of what the connected automobile revolution presents. We believe a huge part of this AI-driven future for connected cars will be more about software and services than hardware. They will be the key differentiators.”


Updates from around the world