Artificial Intelligence (AI), once the bastion of advanced IT teams only, is rapidly seeping into our everyday experiences. From credit decisioning to movie recommendations to checkout-free retail, chances are your life is touched by AI every day, whether you know it or not.
But AI is not just for consumer use cases – it’s also now being used in the fleet management field. Here’s how.
First, a brief explanation of AI models – algorithmic mathematical models that are used for decision making, including making highly accurate predictions. AI models are much more complex than the linear models fleets once used (for instance, calculating oil change dates based on mileage). They can take multiple input variables into account to calculate very specific output, very quickly. AI models get smarter and more accurate over time as they are fed more and more data.
At a high level, AI can be used in two ways: to augment human decision making with new insights and recommendations, and to automate decision making (removing the human component). We’ll explore both.
Today, people still make decisions when it comes to managing fleets, with a lot of assistance from data – for example, driving behavior data, hours-of-service data, environmental conditions, and smart-city data such as traffic and micro-weather conditions.
Many fleets are using AI-based data mining tools to identify outliers, or to discover relationships among different types of data to gain a better understanding of certain behaviors or outcomes. Data mining enables analysis of very large data sets, too large for people to analyze on their own. Equipped with this data, people can then consider business context and make decisions.
An example might be an AI model that determines driver performance based on a series of events over time, and then creates a driver scorecard that rates the driver’s safety. Using AI, the model may provide scorecard details such as, “This driver tends to harsh brake in poor weather conditions.” A manager might then recommend that the driver undergo remedial training for driving in poor weather.
Another example where AI can help augment fleet managers’ decision making is being able to understand fuel economy for your fleet. Fuel is the second biggest operational cost for a fleet, after drivers. AI can help fleets see relationships and make correlations between a specific driver in a specific vehicle on a specific terrain. Using AI, you can see whether it’s the driver that’s the problem, or is it the wrong vehicle for the terrain, or both. Once you know the correlation, you can change the variables to get a better outcome – in this case, better fuel economy.
The bottom line is that by augmenting decision making with AI, managers can make better decisions without putting in a lot of extra effort.
AI can assist with human decision making – or help automate decision making entirely, to eliminate mundane tasks or simply save time.
To build on the example above, AI can not only help create driver scorecards, but also evaluate those scorecards to see where drivers may need help, and then automatically provide it. Instead of waiting for a manager to intervene, a system could automatically send the driver a link to a remedial training course on driving in poor weather conditions. The system could send periodic reminders until the driver registers for and completes the course, and even take the driver off the schedule, re-assigning jobs to other drivers, if the person does not comply.
AI-driven automatic decision making is also becoming more common for fleet maintenance issues. Many fleets today practice preventive maintenance – for instance, changing vehicle oil every 10,000 miles to prevent engine problems. But by analyzing engine data, fleets can get even more precise, and even predictive. Not every vehicle needs its oil changed at 10,000 miles – where, how and how frequently a vehicle is driven all factor into the condition of the oil. Telematics data pulled from the engine and analyzed can give a more precise picture and enable fleets to predict exactly when maintenance is needed.
Fleets can build AI models and workflows that automatically analyze engine data periodically, assess whether action is needed, schedule maintenance as needed, and take vehicles out of service for that maintenance. The workflow could even include automatically ordering parts and supplies.
You can see how such a process would lead to greater efficiency, vehicle uptime and productivity. Preventive fleet maintenance is where AI really shines.
Another feature of AI that could help automate operations is facial recognition. In-cab cameras are a big growth area in telematics. According to Frost & Sullivan, the market for video telematics will grow by 22.2% from 2020-2025, to 3.2 million subscribers. Video telematics can reduce collisions by 60%, and can reduce collision costs by 75%. Visual evidence collected by in-cab cameras is a powerful tool for insurance claims and driver safety training.
These in-cab cameras are not just for looking back on incidents, but can also provide important proactive safety benefits – models that can detect a driver falling asleep or an oncoming accident can trigger audible in-cab warnings that can help prevent a crash. Machine vision technology – powered by AI – can also be used to streamline driver ID and security processes. For instance, you could use facial recognition to unlock or start vehicles, or enable auto-login to your telematics system.
Most fleets don’t have the resources to find and hire data scientists to build their models. Fortunately, vendors are starting to incorporate AI features into their platforms, like a data scientist in a box. This approach masks the complexity of AI, enabling fleets to take advantage of the technology without needing advanced knowledge of AI.
Integrations between telematics vendors and vehicle OEMs are key. These integrations give telematics systems access to rich new data streams for fleets to capture and analyze, for enhanced insight and understanding that again will lead to better performance and lower costs. The growing use of 5G networks will be an enabler, allowing fleets to capture and convey large data sets over the air.
The more data you have, the smarter you can be about making decisions that improve performance, efficiency and most of all safety. Generating actionable insights will be the key to success in this brave new world of AI-enabled fleet management.
Jonathan Bates is head of global marketing for MiX Telematics. Bates joined MiX in 2011 from Europe’s top-selling commercial vehicle OEM, where he held senior positions in sales, marketing, customer success, operations and channel partner development. This article was published under the editorial standards of HDT’s editors to provide useful information to our readers.