top of page

Are Cameras and AI the Solution to Trucking’s Rising Fatalities and Skyrocketing Insurance Rates?

From 2017 to 2018, the number of large trucks in a fatal crash with a weight rating between 10,001 and 14,000 pounds increased 4.6%. At the same time, the number of trucks over 26,000 pounds involved in fatal crashes increased 1.6%.

Meanwhile, from 2016 to 2018, the number of large-truck occupant fatalities went from 815 to 885, an increase of 8%. Of the large truck drivers who were killed by being ejected from their vehicles, more than 40% had neglected to secure their seatbelts, meaning the fatalities were eminently preventable.

What’s behind these horrifying trends? One factor is the ongoing driver shortage, which leads to challenges in both recruiting and training. The shortage also puts added pressure on fleets to get farther faster, with fewer sets of hands on the wheel. Another factor is the still-growing popularity of mobile devices—with the attendant temptation to distracted driving (which affects both the truck drivers and the drivers of other vehicles on the road.)

As if increasing numbers of fatalities weren’t bad enough on their own, the downstream effects on insurance rates are making it next to impossible for many fleets to remain in operation. From 2012 till today, the average verdict against a trucking company has shot up from $2.6 million to more than $17.5 million. Insurance premium costs per mile have gone up more than 17 percent since 2013. The rates rose 12 percent between 2017 and 2018 alone.

While much of the increases can be attributed to unscrupulous litigation, the solution for individual trucking companies has to begin with an impeccable safety record.

That’s where cameras and AI-powered telematics come in.

Tracking Performance and Improving Safety Records

As of early March, trucking demand had gone up 18% to keep supplies on shelves during the coronavirus crisis. Nearly everywhere they go, truckers are being met with crowds expressing their gratitude. It’s difficult to keep count of how many articles are now referring to truck drivers as the “unsung heroes” of the Covid-19 pandemic.

With any luck, this positive sentiment will help clear the way for tort reform making it harder for unscrupulous lawyers to win bankruptcy-inducing settlements against freight companies. But lobbying for such reform will be only one front in the battle to make running a trucking business affordable again.

To achieve an impeccable record, each company will have to take initiative in fostering its own culture of safety. This task can be broken into four parts:

1. Policy

2. Training

3. Tracking

4. Immediate Action

The policy is obviously set by the company’s leadership. But to translate rules into norms adhered to by everyone in the company takes a lot of painstaking effort that begins not just on the drivers’ first days on the job, but even farther back—at recruitment.

Asking questions like, “What motivates you to be as safe as you possibly can be on the road?” are a good start. Counterintuitive as it may sound, answers like “I want to do what’s best for the company and make my boss happy” can be taken as red flags. What happens when the boss is complaining about late deliveries? A driver whose biggest motivation is to make the boss happy would be likely to pay less attention to safety than to speed.

This is why it’s important that both the policy the drivers learn and the training they go through work together to unequivocally convey to them the priority of safety. And so must the incentive structure they work within every day on the road.

Okay, sounds great, right? But how do you train drivers to be safe? And how do you incentivize them to continue improving on their safety practices after their training is complete? This where telematics and performance tracking come in.

There are two simple scenarios where having cameras on a truck reduce your liability. The first is one in which a driver has better visibility into what’s going on around the truck, which makes it easier to avoid accidents. The second is when the camera catches an incident on film and shows that the driver of the other vehicle was at fault.

But a better, more comprehensive approach to driver safety has the cameras trained on the drivers and the areas around the vehicle whenever they’re on the road. It’s one thing to tell a driver it’s against policy to use a phone to text while driving—it’s another thing entirely to show that driver in-cab footage of a violation. Of course, if the same kind of tracking is incorporated into driver training, those types of violation need never occur in the first place.

Behaviors like following too close to the vehicle ahead, distracted driving, cutting turns too short, hard breaking, not buckling your seatbelt, rolling through stop signs—all of these can be tracked using telematics technologies. Ideally, they should all be curtailed in the training process, so by the time the drivers are on the road, they’re well-aware of what they’re doing, and well-aware that their managers are aware.

You’ll usually have better results if you incorporate positive incentives into your training and tracking programs. This will help you avoid an adversarial relationship with your drivers, who are often all too ready to view tracking as spying and telematics as a weapon managers will be naturally apt to use against their workers. Instead, you can award points for compliance, encourage drivers to compete with one another to achieve higher scores, and perhaps even direct some of your insurance savings to the winners. This practice is called “gamification.”

So, you’ve got cameras recording what’s going on in and around your trucks—but who’s watching the feeds all day?

Pattern Recognition and Artificial Intelligence

Well, you can either hire a supervisor to watch feeds all day, or you can program a computer to track what’s going on with your trucks and notify you when an event occurs. But how does the computer know how to interpret what the cameras are recording? How does it distinguish between, say, a driver lifting a sandwich to his mouth and one lifting a phone to dictate a text message?

Some event triggers are relatively easy to program. The velocity the vehicle is traveling when it makes a turn of a certain angle can trigger an alert that the driver is going too fast. Others are more difficult, like distinguishing between a cheek scratch and holding up a phone.

AI addresses these difficulties by training the program to make finer and finer distinctions. You first model the behavior you want to flag to train the computer to recognize the basic pattern, and then you evaluate each instance it flags by referencing the raw footage. The program “remembers” each call it makes by saving the key details that separate good calls from bad. Over time, the alarms will get more and more reliable.

Eventually, you’ll have a tracking system in place that allows you to go about your business until receiving a notification that one of your drivers is engaging in unsafe behavior. You could then pull up the feed for a quick check, and then immediately notify the driver.

You won’t be watching your drivers every minute they’re on the road. But the effects will be the same. With time, your record will reflect the outcome, as fewer incidents occur. And, since the dangerous behaviors are precursors to accidents, your overall safety record will continually improve. That means your risk profile will be lowered, and with it your insurance rates. All the while, the likelihood of a terrible collision—and a company-busting verdict—will be minimized.

Follow us on:

96 views0 comments


bottom of page