Telematics for the transportation and logistics industries has become a vitally important and growing business in the past decade. While Qualcomm and Omnitracs should perhaps be given credit for pioneering the deployment of data-gathering devices in commercial vehicles, since the ELD mandate, numerous telematics service providers (TSPs) have come out of the woodwork. OEMs like Navistar, Volvo, and Wabash National are also very active in putting more technology on trucks and trailers, whether they’re trying to implement predictive maintenance, monitor temperature control, or simply use GPS data to facilitate load matching and route planning.
The result has been a sometimes confusing array of hardware devices that often don’t talk to each other very well. It became clear in May at Transparency18 that the way forward was to develop data standards for interoperability, and that firms building hardware-agnostic software were best poised to win the space.
Lightmetrics, a video telematics software company founded in 2015 by a group of Nokia engineers, is such a startup. The company’s founders, including CEO Soumik Ukil and Product Manager Krishna AG, worked on computer vision, machine learning, and artificial intelligence at Nokia Research. The telematics solution built by Lightmetrics, RideView for fleet managers and insurance companies, is designed mostly around driver behavior, including analyzing high g brake events and reminding drivers of speed limits. Lightmetrics recently closed a funding round led by a Sinapore-based VC firm, and Ukil said that as a result, the company is well-capitalized.
FreightWaves spoke with Ukil and Krishna by phone and they explained some of the points of differentiation that make Lightmetrics unique. The first is Lightmetrics’ go-to-market strategy: because the team has a technology and research and development background, they decided that it made more sense to view TSPs as their clients, rather than approaching individual fleets about their product. Lightmetrics has already been successful with that strategy, and its software is white-labeled by several of the most important telematics service providers.
Secondly, the Lightmetrics team made an important marketing decision that also affected their technology and ultimately made it more flexible. Looking for a way to offer video telematics solutions at multiple price points, Lightmetrics decided to create hardware-agnostic software that could work with any camera, including a driver’s smartphone.
Those two decisions—to build a truly plug and play, hardware-agnostic platform and sell it directly to telematics service providers—in our view, put Lightmetrics in the position to grow rapidly and gain market share.
“The most serious infractions in driving are moving violations like running through stop signs or tailgating, and we realized that all these things have dire consequences, yet fleets had no visibility into them,” said Krishna AG. While other video telematics companies used video cameras as recorders to establish liability in the case of an accident, the Lightmetrics team wanted to leverage their computer vision expertise to allow the camera to truly be another pair of eyes on the road, reading traffic signage in real-time and evaluating a driver’s maneuvers on the road.
Soumik Ukil said that Lightmetrics wanted to build a driver behavior platform that wouldn’t feel like an intrusive monitor that only alerted fleets when bad things happened, but could also help fleets reward safe, efficient drivers.
“Drivers do amazing things to avoid accidents,” Ukil said. “Sometimes we see a swerve or a sudden lane change, and it’s actually to avoid an accident. We’re developing software to detect positive maneuvers, so that when the driver does something great, he or she can be recognized.”
We wondered why Lightmetrics needed to code complex machine learning algorithms to get a camera to read road signs when speed limit data are already present in Google Maps.
“There are two main reasons,” said Krishna AG. “The first is that enterprise map APIs are not exactly cheap; when you already have the camera and computer vision software, reading road signs is actually one of the easier things to do. More importantly, there is no way mapping can be as current as reading the signs with a camera, for example when roadside construction causes speed limits to be revised downward. Washington and other states are moving to variable speed signs,” Krishna continued, “so something that’s live is more useful.”
What to do with all of this data? One of the trends in telematics and the internet of things is ‘edge computing’, where instead of a device on a commercial vehicle uploading huge quantities of raw data to a data center or the cloud, where it then has to be processed and analyzed later, the data is analyzed locally, on the device/vehicle itself, and only actionable insights are transmitted wirelessly to fleet managers. Lightmetrics designed its software to be flexible enough to work with smaller fleets who don’t have data professionals by generating snapshots of specific metrics and basic reports, as well as the larger fleets who might want to do their own API calls to a cloud-hosted database.
“Our goal is to save fleets time by giving them only actionable information,” said Krishna AG, “so we have a fleet manager snapshot: ‘what are the key events and violations happening in the fleet in the last week? Have they improved or gotten worse week over week?’ And fleet managers can look at how violations are trending over a long time, and, for example, the five drivers who need the most attention, and the five who are doing a phenomenal job.”