Fleets have data coming out of their ears – data on fuel consumption, driver behavior, maintenance, compliance, and the list goes on and on. It can be overwhelming to the point they don’t know what to do with it, turning what was supposed to optimize operations into a time-consuming, and ultimately monetary, hindrance.
That’s where Prophecta comes in – at least on the fuel data front.
The new Silicon Valley-based artificial intelligence startup, established at the end of 2022, recently rolled out its vehicle tracking algorithm that aims to reduce carbon emissions by correcting and incentivizing cost-conscious and safe truck driver behavior.
The tool, which operates via app on a driver’s phone or in-cab tablet, serves as an extension of the value proposition of existing telematics systems like Geotab, Samsara and others, said Prophecta Chief Technology Officer Anatoly Volkhover.
“Where our system comes into play is that for every vehicle, we have what we call a digital twin, essentially a mathematical model of how the vehicle performs,” he said. “It's very similar to what NASA does for the space shuttle. When any space vehicle flies, there is a certain level of mathematics that computes its trajectory so you can know what the impact on the trajectory will be if you make a 10-second fuel burn, for example. We do the same thing for a truck.”
The tool corrects driver behavior with real-time coaching for optimal performance based on RPM, load and terrain data, among other things.
“The model is what's lacking in the purely telematics solutions that just collect the data. They give you the data, but you don't know if (drivers) are performing well or not,” Volkhover said. “There is plenty of data that's available today … but most of the data requires human analysis, and of course no one ever has the time to do that in detail. It also requires an education of a data scientist to make sense of it. So the AI that we created is essentially a data scientist in the box. It does that work in real time, and there is no limit to how many data streams it can process.”
