High-tech device allows companies to monitor their drivers

By Jennifer Van Grove

San Diego Union-Tribune

San Diego company Netradyne is on the cusp of launching a technology system that will either act as friend or foe to the 6 million people who operate commercial vehicles in the U.S.

The fledgling firm has invented a device that’s mounted behind the rear-view mirror of a pickup truck, tractor-trailer or van. It acts like a computerized command center, perpetually capturing 360-degree video of the car’s surroundings, automatically identifying everything on the road — from other cars and pedestrians to red lights.

The device also uses artificial intelligence to determine unsafe actions (tailgating, speeding relative to other cars, accidents etc.) and alert managers in near real-time for urgent events.

“It’s all with the purpose of trying to apply AI to revolutionize safety for commercial fleets,” said Sandeep Pandya, president of Netradyne.

The “Driveri” device relies on a Nivdia mobile processor for computing tasks and a Qualcomm chip — the same one used in Apple’s iPhone 5 — for cellular needs, say transmitting footage over 4G.

The device can store five driving days worth of video at a time. It works with a cloud-based management center, where footage can be saved longer term and fleet managers can see alerts, dive into driver analytics and keep tabs on an individual’s daily performance, which is scored at all times by a ranking system.

The goal is safety — for fleet personnel as well as the public — but the implications are grander. The system is likely to influence how insurers determine fault for accidents, thus protecting or exposing drivers and their companies in all manner of incidents.

Netradyne insists it has drivers’ best interests in mind.

“The set of incentives associated with good driving are worth having some technology that basically back the driver up,” Pandya said. “A lot of the time, drivers are victimized.”

Netradyne’s system is slated for commercial availability this month. Exact pricing varies, but the startup said its using a subscription model that will likely cost customers $1 to $2 per day, per vehicle. To date, the startup has financed operations through a combination of bootstrapping and venture capital. In June, Netradyne secured $16 million in its first funding round.

‘Always-on’ awareness

The company aims to be an intelligent leap forward in the realm of telematics, which blends communication and vehicle technologies, and encompasses things like navigation systems or the car-sharing systems that power Car2Go.

Worldwide, the commercial vehicle telematics market is projected to more than double over a five-year period, growing from a $20 billion industry in 2015 to a $48 billion market by 2020.

Netradyne competes with a number of existing fleet management systems from companies. The startup believes its AI capabilities set its platform apart from those offered by rivals.

“A lot of the legacy providers, what they do is wait until the vehicle stops suddenly. Then, they turn on the camera and record what’s happening around the vehicle,” said Adam Kahn, Netradyne’s vice president of fleet business.

“Someone, somewhere else, maybe a day later, will study that video and try to determine what happened,” Pandya added.

Meanwhile, Netradyne’s device, Kahn said, can manage a trillion calculations per second, which allows for complex interactions with objects on the road. This kind of always-on, computer-vision-based awareness.

Of course, professional drivers might be unsettled by the Driveri system — and reasonably so. Netradyne’s technology can track and record pretty much all behaviors in and around vehicles, meaning drivers are essentially being micromanaged by algorithms. Have a bad hour or two of driving and Netradyne will tell your manager.

The company counters that the commercial fleet industry is swiftly moving toward video technologies, so spy-level monitoring is merely a given. Plus, Netradyne’s system is designed to not just penalize poor performance but reward good behaviors, too.

“Most systems out there, if not all the systems, will focus on the things you’ve done wrong. You were speeding. You ran a red light. You got in a crash. With the utilization of artificial intelligence, we can watch every minute of every driving day,” Kahn said. “If I can watch everything … I can say, ‘Hey, my driver did a good job today.’”