New Diagnostic Metric for Battery Management Systems

Engineers at the University of California, Riverside, have developed a new diagnostic metric called the State of Mission (SOM) designed to predict whether a battery can safely and successfully power a specific task. SOM uses both battery data and environmental factors to generate real-time, task-specific predictions. This hybrid approach combines the strengths of rigid physics equations and machine learning models, allowing for reliable predictions even under stress.

Key Takeaways:

• The SOM diagnostic metric uses a hybrid approach to combine battery data and environmental factors to predict battery performance.
• SOM can generate real-time, task-specific predictions, answering the question of whether a battery can safely and successfully power a specific task.
• The model combines the strengths of rigid physics equations and machine learning models to make reliable predictions.
• SOM was tested using publicly available battery datasets from NASA and Oxford University, showing a significant reduction in prediction errors.
• The model can tell a driver whether they can complete a planned route, but may need to recharge halfway, or that a drone flight is not feasible under certain wind conditions.
• SOM offers a smarter, forward-looking output, transforming abstract battery data into actionable decisions.
• The main limitation of the SOM model is computational complexity, requiring more processing power than current lightweight, embedded battery management systems.
• The researchers plan to test SOM in field environments and expand its capabilities to work with other battery chemistries.

Statistics:

• Prediction errors were reduced by 0.018 volts for voltage, 1.37 degrees Celsius for temperature, and 2.42% for charge state compared to traditional battery diagnostic methods.
• The model demands more processing power than today's lightweight, embedded battery management systems typically provide.
• The researchers plan to test SOM in field environments and expand its capabilities to work with other battery chemistries such as sodium-ion, solid-state, or flow batteries.

Sources:

• Ozkan, et al. "A mission-aware measure for battery-state estimation and prediction of task-specific performance." iScience (2025).
• University of California, Riverside. "Researchers Develop New Diagnostic Metric for Battery Management Systems" (2025).