AI can prevent truck breakdowns

Imagine if you knew when your truck would break down before it happened. This is possible, thanks to artificial intelligence (AI)!

The vast amounts of data being collected from trucks today mean it is increasingly easier to identify and fix faults before they lead to unexpected breakdowns.

Traditionally, the main approach to maximising uptime has been regular, scheduled servicing and reactive measures like break-down support services. But with the range of sensors and wireless technologies typically found on today’s trucks, businesses can now be far more proactive.
 

How can data and wireless technology prevent breakdowns?

At the heart of connected services and preventative maintenance is the fact that using wireless technology and sensors now makes it possible to collect vast amounts of data from a vehicle in real-time. By analysing this data and identifying patterns, it is possible to predict and anticipate a fault before it occurs. This gives you time to schedule a workshop visit at your convenience, and fix the fault before it causes an unexpected breakdown.

“In the short time I have been working in this field, I have seen the technologies and our capabilities expand exponentially,” says Matthias Tytgat, manager of Volvo Trucks’ Monitoring Centre in Ghent, Belgium.

“In 2016, we were remotely monitoring just one component and it took us a whole day to complete a full check in a fleet of several hundred trucks. Today, we’re monitoring multiple components in tens of thousands of trucks, and we can complete a full check of the whole fleet in just eight minutes. The exciting part is that we’re constantly improving,” he continues.
 

The role of artificial intelligence in reshaping the truck industry

The more data a system can analyse, the more accurately it can predict outcomes. Initially, connected services and real-time monitoring services were designed to predict faults by reacting to certain thresholds or sensor values for individual parameters (for example, the engine exceeding a set temperature).

“While these sorts of insights are useful, this can be somewhat limited because it does not take into account the vehicle’s unique circumstances and driving conditions,” explains Tytgat. “While it’s important to detect a potential fault as early as possible, it is also important not to bring a vehicle into the workshop unnecessarily.”

Machine learning can be used to analyse greater volumes of data and to detect patterns impossible to define by a normal set of rules. This results in even more accurate predictions; different parameters and data points from a wider variety of components and sensors can be combined. AI systems then analyse these combinations to detect patterns indicative of potentially problematic behaviour likely to lead to a breakdown.

For example, the temperatures of different parts can be analysed in combination with other factors such as vehicle mileage and fault codes. Once a machine learning algorithm has been trained to identify a pattern or combination of factors that often cause a particular fault, it becomes possible to predict problems for individual vehicles no matter what type of application they’re in.

“It will be as if the service was created for a specific vehicle and its customer,” says Tytgat. “And as we continue to improve our capacity to analyse data, the more accurate these systems will be.” It sounds as though the future is going to be a great place to be.