Commercial vehicle operators and suppliers are learning to love advanced technologies, including AI, because they offer the ability to improve efficiency and cut costs. Steve Banner looks at the effect AI is having on telematics
(Image credit: AdobeStock by Azhorov)
The term AI – artificial intelligence – is fast becoming one of the most abused and misused in the English language. Marketing departments delight in throwing it around when what they are really talking about is intelligent manipulation of data; something that has been around for many years.

“A distinction has to be drawn between weak and narrow AI and strong AI,” says Michelin Connected Fleet marketing manager, Matt Childs. The former requires human input, the latter describes a situation where a machine rather than a human being is in sole charge of a particular activity. “The transport industry is still at the former stage, with AI involving machine learning and the collection and processing of data,” he observes.

“And fleets are having to deal with huge amounts of data produced by vehicles,” points out Raghunadh Banerjee, vice president, data solutions, at Bridgestone Mobility Solutions. “It’s coming from TPMS (tyre pressure monitoring systems), cameras, sensors, refrigeration units and a variety of other sources.”

Specifically, it is the speed at which AI can analyse information that is starting to deliver palpable benefits to fleet managers, says Barney Goffer, UK product manager at Teletrac Navman. That is especially the case when it comes to reviewing footage from cameras mounted in truck cabs that are trained on the driver and on the highway ahead, he contends.

The driver may be looking at his smartphone, he may be otherwise distracted, he may be drinking or smoking, and he may be driving too close to the vehicle in front. If a manager is responsible for 200 trucks with 200 inward- and outward-facing cameras, then it will take him forever to analyse all the footage and spot the most dangerous kinds of behaviour.

AI can do so almost instantly, however, and highlight the top half-a-dozen undesirable activities that require prompt action. “It makes it easier to identify risk,” Goffer states.

You can quickly see which of your drivers are the best, and which are the worst, when measured against your safety KPIs (key performance indicators), says Banerjee. The worst ones are likely to damage your trucks which means your insurance costs will go up, he says; and that‘s the best-case scenario. The worst case is a major accident with multiple fatalities. As well as informing fleet managers about what is happening in real time, AI can give drivers an immediate heads-up, too. “They can be told ‘you’re not paying attention’,” he says.

Accidents may be avoided as a consequence. Asks Inseego UK managing director, Steve Thomas: “Why fit a camera that simply films an accident when you can fit a camera that prevents one?"

If you run a fleet of, say, 300 vans, then you can bet that every driver will exceed the speed limit by 1mph to 2mph from time to time, he observes. “That’s not the same as hurtling past a primary school mid-afternoon when all the kids are coming out though,” Thomas says; the sort of reckless behaviour that AI can spotlight.

“In the same way, taking a quick sip of water while you are at the wheel is not as bad as staring at your smartphone for minutes on end,” he remarks. AI can be used to differentiate between the two activities.

Adds Goffer: “It can contextualise an event. Harsh braking may have occurred because somebody pulled out in front of a driver and he needed to brake heavily to avoid a collision; so that‘s good driving.“

It can be used to combine different data sets, says Thomas, and identify heightened risk levels as a consequence. “AI can spot if a driver has been driving for four hours without a break in wet weather and happens to have an endorsement or two on his licence,” he observes. Add up all these factors, and the danger level escalates; so maybe he should be advised to pull over whenever it is safe to do so, and take a break.

The data gleaned can be used to tailor training to the individual driver. “If it shows that he regularly brakes harshly without a good reason then this can be addressed with an online training course,” he says.

AI can also assist with route planning and scheduling, Banerjee contends. Cautions Thomas: “Remember, though, that a lot of what we are talking about here is the manipulation and integration of data and that‘s been going on for quite some time.” 

The same can be said about figuring out how many chargers you are likely to need if you are introducing electric vehicles and where they should be located, Thomas adds.

Remarks Childs: “Saying that AI can assist with route planning is rather overstating it. What‘s really involved here is the overlaying of a lot of data sets.” 

AI can certainly help fleets introduce a preventive maintenance schedule if they have not already done so, says Banerjee. If a particular make and model of truck has a key component fail after so many thousand kilometres – and this has happened a number of times – then AI can spot this trend and alert the fleet manager accordingly. He or she can then ensure that all the vehicles likely to be affected are booked into a workshop in good time to have the part concerned replaced, and avoid ending up with a truck broken down at the roadside. “AI can also be used to predict when tyres are likely to need replacing,” he observes.

Adds Childs: “You can soon see which of your trucks are costing you the most to maintain.” 

AI can play a vital role when it comes to identify which delivery routes can be serviced using electric trucks, Banerjee believes. “If you’ve got, say, 20 covering less than 200km daily, then it can quickly spot which ones they are,” he says; something which may not be immediately obvious in a big fleet. “It helps you plan your future sustainability,” he remarks. 

AI chatbots can provide invaluable assistance, Banerjee contends. “You don’t need to be a technical expert. All you need to say is ‘Hey Alexa, which are my worst-performing trucks?’ and it will tell you.”

AI already has a role to play so far as the burgeoning requirements of London's DVS (Direct Vision Standard) are concerned, says Childs. “It comes into play when cameras need to distinguish between fixed and moving objects,“ he points out.

At present the use of AI is largely the preserve of major fleets, but Banerjee believes this will change over the next two or three years as smaller operators become attuned to its benefits. “There’s a cost to introducing it at present, but the cost will come down,” he predicts.

"AI will end up doing more and more of the work, and the fleet manager will have to do less and less," says Thomas. 

In effect AI is handling much of today‘s drudgery, But might it ultimately result in the fleet manager being replaced entirely as it grows and develops? Nobody is confident enough to make such a prediction, but with driverless trucks on the horizon – and $1bn being pumped into AI-based autonomous vehicle technology start-up Wayve by Microsoft, among others – one has to wonder. And one cannot help but speculate about how traffic commissioners would cope with such a development. 

Related content