When we look at the Battery Management System (BMS) History and Roadmap, such as that shown by Zhang et al [1] we see a long period of simple functions and reporting. Then an exponential growth in capability with Lithium Ion chemistry, the cloud and machine learning.
The four generations of BMS are described by Zhang et al as:
| Generation | Functional Characteristics | Application Area |
|---|---|---|
| First | Basic functional | Early electric tools, lead-acid battery systems, low-end EVs |
| Second | Digitalization and algorithms | HEVs and early EVs |
| Third | Intelligent and integrated | High-end EVs and energy storage systems |
| Fourth | Global collaboration and AI-driven | Next-generation intelligent EVs and large-scale energy storage |
The Automotive Council UK & APC Technology Roadmap 2024 [2] describes BMS capability that includes:
- upgrades using machine learning on historic data
- BMS capable of physics-based SoH and SoC cell tracking
- AI enabled BMS using self-updating algorithms and in-situ data
So, let us look at the 4th Gen BMS, where is is going and what to expect.
Fourth Generation
The fourth generation takes advantage of the cloud based data, machine learning and the exponential growth in sensing development and cost reductions.
1. High-Precision Sensing
Where it is going: from basic voltage and temperature to multi-modal internal sensing:
- fibre optic
- ultrasonic
- strain
- pressure
- chemical signatures
What to expect: internal sensors become standard on premium EVs. Early detection of dendrites, micro-cracks, SEI issues, thermal anomalies. Safety moves from reaction to prediction.
2. AI-Enhanced SOC, SOH, SOP
Where it is going: Hybrid estimation replaces pure EKF. Models plus ML, LSTM, CNN spectrograms, EIS + deep learning.
What to expect: Adaptive algorithms learn from user habits, climate, charging, aging. SOC error below one percent. SOH becomes a live variable instead of periodic labs.
Why Physics-Informed AI is the Future of BMS
Developing a generalized AI model for State of Health (SOH) prediction is a “boss level” challenge in battery management.
While standard Neural Networks (LSTMs/GRUs) are excellent at capturing temporal patterns for a specific batch, they often fail to generalize across diverse chemistries (LFP, NMC, NCA), variable C-rates, and fluctuating temperatures. They lack “physical common sense”, sometimes even predicting that a battery’s health can “heal” overnight.
This gap can be filled, by moving toward Physics-Informed Neural Networks (PINNs).
Image credit: Ma, Hongli, Xinyuan Bao, António Lopes, Liping Chen, Guoquan Liu, and Min Zhu. 2024. “State-of-Charge Estimation of Lithium-Ion Battery Based on Convolutional Neural Network Combined with Unscented Kalman Filter” Batteries 10, no. 6: 198.
3. Predictive Safety
Where it is going: beyond voltage alarms. DL predicts thermal runaway 8 to 13 minutes early.
What to expect: predictive safety becomes mandatory for HV EVs. Real-time anomaly detection in cloud fleets. Preventive actions before drivers notice anything.
Basab Ranjan Das Goswami et al [3] show how a graph neural network (GNN) for spatial change is coupled with a Long Short-Term Memory (LSTM) network are used to predict battery temperatures in a module based on spatial and temporal temperature data obtained from temperature sensors. One approach that can be used to predict thermal runaway earlier.
4. Active and Intelligent Balancing
Where it is going: shift from passive resistive balancing to active bidirectional solutions. Efficiency more critical as packs age.
What to expect: Power-electronics balancing at cell-group level. Thermal plus electrical balancing combined. Range stability becomes a key OEM KPI.
5. High-Integration BMS Chips
Where it is going: BMS controller, charger controller, and sometimes inverter integrated into one unit.
What to expect: AI accelerator chips inside the BMS. Wireless BMS for weight reduction. One unified brain managing the full e-powertrain and safety envelope.
6. Edge-Cloud Collaboration
Where it is going: onboard BMS keeps real-time protection. Cloud handles long-term learning and analytics.
What to expect: digital twins for each pack. Cloud-based RUL prediction, charging limits, fast-charging constraints, fleet-wide degradation modelling. Cloud becomes the battery life extender.
Why is Remaining Useful Life (RUL) important?
With the ability to predict the RUL accurately, the users, OEM’s, fleet operators or insurance companies can have a better understanding of the time at which the battery will reach its EOL and, based on that, plan any maintenance activities proactively. Accurate RUL also plays a crucial role in the second-hand car market, helping end-users gain trust and confidence in the battery pack’s longevity and enables fleet operators to plan the redeployment of used packs in second-life applications.
7. Full Lifecycle Management
Where it is going: From pack protection to managing charging, second-life, recycling.
What to expect: BMS data drives resale value, second-life grading, warranties, safe disassembly. Battery lifecycle becomes a monetizable data product.
The new EU Battery Regulation 2023/1542 entered into force on 17 August 2023 and covers the whole lifecycle of batteries from production to reuse and recycling. While the Battery Regulation is already in force, further legal documents will be published in the coming years specifying certain aspects of the implementation. Among the new requirements, for example, is the Battery Passport that goes with every new LMT, industrial (> 2 kWh), and EV battery and contains material pathways, product specifications and lifetime data.
References
- Zhang, Q.; Shang, Y.; Li, Y.; Zhu, R. A Concise Review of Power Batteries and Battery Management Systems for Electric and Hybrid Vehicles. Energies 2025, 18, 3750
- Electric Energy Storage Technology Roadmap 2024, Automotive Council UK & Advanced Propulsion Centre
- Basab Ranjan Das Goswami, Yasaman Abdisobbouhi, Hui Du, Farzad Mashayek, Todd A. Kingston, Vitaliy Yurkiv, Advancing battery safety: Integrating multiphysics and machine learning for thermal runaway prediction in lithium-ion battery module, Journal of Power Sources, Volume 614, 2024