AI is making waves across various industries, and electric vehicles (EVs) are no exception. While the technology comes with a bunch of promising advantages, it also has a darker side that could pose risks to both drivers and passengers.
One of the most exciting uses of AI in EVs is in the battery estimation system, which aims to make vehicles runs more efficiently. Knowing exactly how much charge is left in the battery isn’t just helpful; it’s crucial for keeping passengers safe on the road.
Understanding Battery Charge Levels
The State of Charge (SOC) is a handy number that shows how much juice a car’s battery has left, usually expressed as a percentage. This metric is vital as it helps to avoid issues like overcharging or running the battery completely flat.
If SOC isn’t measured accurately, it could lead to serious problems, including battery fires caused by overheating, a nightmare scenario known as “thermal runaway.”
It’s this potential danger that really makes the electric vehicle industry wary about fully embracing AI, despite its efficiency boosting capabilities.
Why AI Adoption is Stuck
According to Martin Skoglund from the Research Institutes of Sweden (RISE), the real reason AI is struggling to get its moment in the spotlight in the EV space is due to trust issues.
“Previously, the fear was that AI components aren’t reliable for vehicles since they rely heavily on data.Causality and liability become murky because AI operates like a black box,” he shared in an interview with Euronews Next.
He added, “It’s a challenge because you can’t look inside and figure out what went wrong.”
On the flip side, conventional systems work with precise calculations that can efficiently track metrics like current and voltage, making them reliable and safe.
Shifting Perspectives on AI
Interestingly, the narrative around AI is evolving. Recent advancements have led researchers to train AI on extensive battery data to pick up subtle signs of voltage, current, temperature, and other variables.
“What changes the landscape is that battery technology is advancing at a lightning fast pace. The traditional methods can’t keep up and often miss out on important details like battery aging, which AI could potentially catch,” Skoglund explained.
The Pros and Cons of Incorporating AI
Incorporating AI into battery systems could mean that EVs can extend their range and lifespan, but it’s a double-edged sword that could endanger lives if executed poorly.
To better understand these risks, Skoglund and his team conducted tests known as “fault injection experiments,” which mimic potential software malfunctions.
They found that even the tiniest errors in the AI’s input could lead to huge discrepancies in what it reports, resulting in charge levels that don’t reflect reality at all.
Such inaccuracies could put lives at risk, creating the possibility of abrupt vehicle stoppages or battery overload scenarios, which could trigger fires.
Finding a Middle Ground
As a solution, researchers are proposing a “safety cage”—a thoroughly vetted system to monitor the AI’s performance in EVs by running simple checks on parameters like voltage, current, and temperature.
If the AI goes haywire, the safety cage would automatically shut it down. This hybrid model offers a blend between the traditional safety measures and cutting-edge AI technology, paving the way for a safer future in electric vehicles.
