Telematics can tell fleet managers a lot about how drivers are performing on the road, like when (and how many) harsh braking, hard cornering, and speeding events occur. Add outward-facing video, and fleets can see what’s going on in front of the vehicle at the time.
While together telematics and outward-facing video can give a fleet manager a lot of great data to work with, there’s still part of the picture missing.
Take, for instance, one driver who appeared to be the best in the fleet. Based on the data points available, his performance was stellar. He was always on time, never missed work, and never got ticketed.
But when the company installed smart video, which uses inward-facing cameras, artificial intelligence, and machine learning to capture instances of unsafe driver behavior, a different picture emerged. The driver was dozing off more than five times per night — while driving at highway speeds.
“He was suffering from sleep apnea and was a ticking time bomb for a major accident,” said David Isler, CEO of FleetCam. “Our customer coached him, and he got the help he needed to stay alert and safe on the road. This is just one example of using new smart video data points to get ahead of potential problems.”
Smart video is giving fleet managers a virtual ride-along, without weeding through endless hours of footage, and it’s helping fleets become safer and more efficient. Here’s how.
What is Smart Video?
First things first: What exactly is smart video?
Smart video systems incorporate additional cameras beyond traditional dashcams, allowing fleets to see both inside and sometime around the vehicle. But beyond capturing more footage, artificial intelligence (AI) identifies specific driver behaviors and helps to pinpoint the most critical moments from that footage so fleet managers can view the parts that yield the most valuable information.
Sid Nair, senior director of transport and compliance at Teletrac Navman, explained to understand smart video, you need to understand the history of video telematics. Nair said video evolved in three distinct phases: the first was pure recording, primarily used for accident reconstruction.
“It required you, as the driver in the truck, to know what happened at what time and then provide that information to your fleet manager to find the right video footage,” he said.
Nair called Phase II “augmented intelligence video,” which allowed video to stream to a back office where people evaluate it.
“Augmented video has some triggers for events like harsh braking and flags the moment for people to look at the video stream,” Nair said. “In Phase III, we evolve into smart video, which allows us to detect actual driver behavior and stitch it all together. Now we can get a 360-degree view. The smart part of it comes when you can connect the dots between looking at the road and the driver.”
Let’s say a forward-facing camera shows a car cutting off one of your drivers. The telematics device detects harsh braking. But whose fault is it? Is it the car’s fault for cutting off your driver? Or was the driver looking at his phone, causing the harsh braking?
Inward facing cameras provide the answer by detecting the driver’s exact behavior and alerting the fleet manager.
“The camera itself can now detect these things rather than it happening in the back office,” Nair said. “Smart video looks at what happened and why then flags it as needed. Fleet managers don’t have to spend time going through footage.”
Jason Palmer, COO of SmartDrive, said smart video adds a new level of analysis that provides actionable information to both drivers and fleet managers.
“Initially, video telematics was combining video information with breadcrumbs, such as events on a map, allowing you only to see the video and know where it was recorded. Now, video telematics allows fleets to bring in the information and integrate it with back-office workflows,” Palmer said. “As opposed to less sophisticated video, which just records what’s occurring, ‘smart video’ records and performs analysis at the same time, alerting the driver as to what the system is seeing, along with providing analytics to the driver’s manager.”
In addition to providing more information about driver behavior and the context of unsafe driving events, smart video also makes it easier for fleets to drill down to the information that matters most, thanks to AI and edge computing.
“Regular video cameras/dashcams merely record what is happening, and you typically would have to search through hours of footage to find a moment or incident that took place,” said Ryan Driscoll, VP of marketing for GPS Insight. “Smart cameras can isolate and transfer important moments and info for real-time notification and easy access. Smart cameras can detect, analyze, and notify management of accidents, hard braking, rapid acceleration, driver drowsiness, traffic-signal violations, stop-sign violations, tailgating, seatbelt compliance, U-turn detection, and of course speeding based on the posted speed limit.”
While smart video does record every minute on the road, AI and edge computing analyze these moments and only notify managers about the moments that matter most.
“Smart video uses AI and machine learning to intelligently capture and automatically classify that video data the dashcam is capturing,” said Kevin Aries, head of global product success for Verizon Connect. “Now, fleet managers can understand how severe an event is, and it helps them focus only on the relevant pieces of data so they can manage safety around their business.”
In addition to correcting driver behavior, smart video can also be used to identify and reward the best drivers.
“We’re capturing every minute of driving time so we can identify not only when a driver is compliant with laws and policies, but when they demonstrate exceptional driving,” said Adam Kahn, president of fleet for Netradyne.
This is especially important when sometimes “unsafe” driver behaviors may be good defensive driving. Smart video is smart enough to distinguish when a behavior is truly positive or negative.
“Video takes telematics to another level by being able to give context and identify positive driver behaviors,” Driscoll said. “For example, if a driver slams the brakes to avoid hitting a child that ran into the street to chase his ball, with regular telematics data, the driver would be dinged for hard braking, and it would negatively affect his score. With video, the AI on the camera would see that the hard braking was to avoid hitting someone so that it would give them a positive score.”
And, perhaps most importantly, smart video can be used to assist drivers who are in trouble. “The biggest difference between video and ‘smart video’ is the ability to deliver data in real-time,” said Ashim Bose, chief data scientist and VP of artificial intelligence, machine learning, and data engineering for Omnitracs. “A connected video device can relay information to fleet managers immediately, which can be critical if a driver is in an accident.”
Isler said that while AI has become increasingly useful, it’s also become more accessible to fleets.
“The price has come down significantly, and at the same time, AI has become much more powerful,” he said. “In the past it would not have been possible to compute simultaneously tracking distance to the vehicle in front of the driver, checking to see if he is smoking or using a cell phone, watching his eyes through sunglasses using infrared to see if he is falling asleep, and recording five-plus channels of video in and around the vehicle. It is amazing when you stop to think about how much is happening at the same time!
Taking a closer look, here’s how the technology behind smart video operates.
“First, a sudden change in driving temperament — for example, harsh braking, rapid acceleration or unexpected lane departure triggers the camera associated with that specific sensor to turn on and capture 3-5 seconds of footage,” said Will Wycks, global marketing director for Chevin. “This reveals the cause of that change.”
How does smart video know which instances to capture? Artificial intelligence and machine learning are essential.
“Artificial intelligence needs a lot of data to train the algorithms to work,” said Michael Phillippi, VP, technology for Lytx. “We have data from 600,000 vehicles, with events we had already classified. We then took that data, fed it into a machine, and out comes a machine algorithm. In other words, years of human judgment about what’s happening in a video is fed into a machine, and then the machine can tell you in the next video what’s happening. That’s how it’s able to find the unique moments in time that are important to fleets and drivers.”
Thanks to AI and machine learning, smart video can remove the potential for human error or bias as well as the time-consuming processes associated with manually reviewing video.
“Video is data, and historically that has been left to the human/viewer to analyze and interpret. That takes time and is based on opinion,” Aries said. “Smart video uses advanced technology to interpret better what’s happening in those videos. Ultimately, it helps inform decision making by being able to tell the viewer what’s happening in the video and how safe or unsafe an event has been. It also saves time by reducing the need to review hours of video footage and helps them identify important events.”
Kahn said the systems that make video “smart” allow for real-time monitoring and improved performance over time — attributes you don’t get with first-generation video telematics.
“Analysis is done on the device with edge computing, resulting in near real-time notification and updates. Legacy systems have a delay in that they upload the data in batches overnight, the video is scored by humans, and returned later,” he said. “Machine learning also enables the system to be trained for consistent precision — and continues to get better over time. Legacy video telematics relies on a human making a trained judgment.”
Phillippi said with continuous distraction detection and cameras always trained on drivers, machine vision and artificial intelligence (MV+AI) can detect things like cell phone use, smoking, eating and drinking, seatbelt use, and camera obstruction.
“We use video, paired with MV+AI, to evaluate how the driver is doing and provide a scorecard associated with it. Then, we provide fleet managers with the highlight reel with the unique moments to share with drivers,” he said. “It’s the best of both worlds — a scorecard that tells you exactly what you were doing and when, and then some unique videos that drive home the message by showing drivers the footage. It’s comprehensive visibility and good teaching moments. Fleets combine those two to drive the risk out of their fleets.”
How Does Smart Video Make Fleets Smarter?
What does this advanced technology do for fleets? It allows them to identify driver behavior they may have never known about, were it not for the footage. Using video footage of the most critical moments of a driver’s trip allows them to provide more effective driver coaching. The result is safer drivers on the road and reduced risk of collisions.
There’s also what other drivers are doing. Smart video can help exonerate truck drivers in the case of collisions and reduce costs associated with liability.
“The two most important benefits are fewer accidents and improved security. Smart video provides context for driver coaching that fleet managers could not otherwise have unless they ride shotgun with all of their drivers,” Driscoll said. “Smart video also captures positive driver behavior and can exonerate drivers from false claims and not-at-fault accidents, which is a layer of security that businesses and drivers have never had before. The result is saved costs due to reduced accidents, insurance premiums, and costs associated with false claims.”
Phillippi said data shows how video solutions, especially those paired with MV+AI, pay off. “The largest benefit we see straightaway is a 50% to 80% reduction in collisions,” he said.
“Larger vehicles have higher per-claim costs per vehicle, so generally, in two to three months, the return on investment pays off.”
Aries noted that smart video builds on the results telematics solutions currently provide fleets.
“According to a survey we did last year, customers that use telematics solutions saw a decrease in accidents up to 22%,” Aries said. “We believe integrated video is an extra layer on that success. Video today has captured hundreds of harsh driving events that managers would never be able to see if it weren’t for the video solution.”
Above all is the added protection smart video offers drivers.
“The number one benefit is adding a level of safety to make sure that the driver gets home to his family at the end of his shift,” Isler said.
How to Handle Two Top Smart Video Challenges
Challenge #1: Driver Skepticism
“Some organizations get pushback on driver-facing cameras around an invasion of privacy. Long-haul drivers even argued that, because they often sleep and eat in their trucks, it’s their home and object to being watched in their home,” said Ryan Driscoll, VP of marketing for GPS Insight. “That argument fell short with the California Attorney General, which concluded that the use of the cameras doesn’t violate any state laws.”
How to Overcome It: “It’s important drivers understand how the program is benefiting him/her and the company, how the video content is being used, etc.,” said Jason Palmer, COO of SmartDrive. “Along with these issues, it’s important to address any privacy concerns.”
Once drivers experience the technology, the benefits become apparent. “The biggest challenge is probably fear of ‘Big Brother,’” he said. “It takes a little time and sometimes even a pilot to get drivers to understand this technology is guarding them and keeping them safe. Once they have driven with it, they understand: All it takes is driving one time while distracted or sleepy to save their life for them to be very thankful it is there.”
Adam Kahn, president of fleet for Netradyne, agreed that once the system is in place, it can ease drivers’ concerns.
“Though video is becoming more pervasive, there is still some driver resistance at first when fleets adopt. We can quickly overcome that notion as drivers experience positive feedback and realize it’s there to protect them and their fleet,” he said.
Challenge #2: Belief in Data Alone
“People think by just throwing a camera in a truck, magic is going to happen. It does not,” said Sid Nair, senior director of transport and compliance at Teletrac Navman. “Smart video still requires a human element to make sure the technology is working for you. I’ve seen many people underutilize technology when there’s no one managing it. Buying a wrench doesn’t fix your car problems. Someone needs to be at the helm.”
How to Overcome It: “Smart video solutions provide the data, but the challenge is on companies to leverage that data to have more informed conversations with drivers,” said Kevin Aries, head of global product success for Verizon Connect. “The companies have to develop the culture and conversation around safety and safe driving. That is ultimately the call to action we’d have for businesses adopting these solutions.”