According to ITNA and citing Interesting Engineering, an 18-year-old student in the United States developed a model that can help solve one of the biggest long-term challenges of electric vehicles: determining the true health of aged batteries.
Colin Jcho, a student at Nueva School, presented a framework for estimating the state of health of lithium-ion batteries with a prediction error of only 2.36 percent.
This project made Jcho a finalist in the Regeneron Science Competition in 2026, which is one of the most prominent science and math competitions in the United States.
It is worth noting that electric vehicle batteries gradually degrade with charging, discharging, and exposure to fluctuating temperatures, but determining the remaining useful life in a battery is not a simple task.
The battery management system cannot simply look inside a cell and directly measure degradation, so engineers must infer its condition from electrical behavior.
According to the Institute of Science, Jcho analyzed data from 22 batteries that had been deliberately aged with electrical signals designed to simulate variable driving behavior, and then developed a model that combines equivalent-circuit modeling with machine learning.
Jcho’s approach did not rely entirely on AI; rather, he combined machine learning with an equivalent-circuit model that represents the battery’s complex electrical behavior using more analytically tractable mathematical components.
Jcho conducted this research through Stanford University’s Young Investigators program and in Stanford’s Energy Control Lab, and the results were later presented at the Modeling, Estimation and Control conference in Chicago and published in the Journal of the Electrochemical Society.
More accurate estimation of battery state of health can improve the safety of electric vehicles, maintenance planning, and battery lifespan.