Engineers Develop Predictive Battery Tool to End Range Anxiety for Electric Vehicle Drivers

Drivers of electric vehicles may soon stop guessing whether their cars can make it home. Engineers at the University of California, Riverside, have developed a new diagnostic system called the “State of Mission” (SOM), designed to predict whether a battery can safely and successfully power a specific journey under real-world conditions. Unlike today's systems that show only a percentage of charge, the SOM model factors in terrain, temperature, and traffic to estimate if the vehicle can finish a planned route without running out of energy.

UC Riverside's Hybrid Battery Model Accurately Predicts EV Range and Real-World Performance

According to a report published in iScience, the research team led by Mihri and Cengiz Ozkan developed SOM by combining data-driven learning with physics-based models. This hybrid approach uses battery charge, discharge, and heat data while following electrochemical and thermodynamic laws, resulting in highly reliable predictions even under changing conditions such as steep climbs or sudden cold weather. This method bridges a gap between simple charge estimates and realistic, mission-aware insights.

In addition, the team trained the model using publicly available datasets from NASA and Oxford University to test its accuracy. This contains long-term data on voltage, temperature, and performance cycles. Results showed SOM could reduce prediction errors to just 0.018 volts for voltage and 1.37°C for temperature—far superior to current diagnostic systems. Unlike a basic charge display, SOM could tell drivers whether to recharge midway or if a drone mission is unsafe due to wind.

Researchers mentioned the model could make electric vehicles, drones, and even grid storage systems safer and more efficient by converting complex battery data into actionable insights. “It transforms abstract battery data into real-world decisions,” Mihri Ozkan denoted in the report, noting that it enhances planning and reliability for all energy-driven technologies. Although the framework still demands high computational power, experts believe further optimisation will make it suitable for commercial EV systems.