Notice of Pre-AIA or AIA Status
The present application, filed on or after March 16, 2013, is being examined under the first inventor to file provisions of the AIA .
Information Disclosure Statement
The information disclosure statement (IDS) submitted on 07/25/2025 & 01/22/2024 are in compliance with the provisions of 37 CFR 1.97. Accordingly, the information disclosure statement is being considered by the examiner.
Claim Rejections - 35 USC § 101
35 U.S.C. 101 reads as follows:
Whoever invents or discovers any new and useful process, machine, manufacture, or composition of matter, or any new and useful improvement thereof, may obtain a patent therefor, subject to the conditions and requirements of this title.
Claims 1–5 & 7-10 are rejected under 35 U.S.C. 101 because the claimed invention is directed to a judicial exception (i.e., an abstract idea) and does not include significantly more.
Claim 1 is representative and recites, in substance:
• collecting time-series data related to operational parameters of a battery pack
• processing the data using a nonlinear state space reconstruction algorithm
• training a first LSTM neural network to predict SOH (440)
• predicting SOH (440) using the first LSTM
• reconstructing a second phase state space
• training a second LSTM neural network using the reconstructed data and estimated SOH (440)
• predicting SOC (430) using the second LSTM
The limitations directed to collecting data, processing data, training machine learning models, predicting values, and feeding outputs of one model into another model collectively recite data analysis, mathematical modeling, and prediction steps.
These limitations constitute the abstract idea. The remaining limitations are additional elements.
Step 1 – Statutory Category
Under Step 1, Claims 1-10 fall within a statutory category: Claims 1–10 recite processes. Accordingly, Claims 1-10 satisfy Step 1.
Step 2A – Prong One - Judicial Exception
Under Step 2A Prong One, claims 1–5 and 7–10 recite an abstract idea under the 2019 Revised Patent Subject Matter Eligibility Guidance.
The steps of nonlinear state space reconstruction, delay embedding, normalization, error evaluation, and neural network training involve mathematical relationships, data transformations, and model optimization, which fall within the mathematical concepts and mental process groupings of the Guidance.
Accordingly, claims 1–5 and 7–10 recite a judicial exception.
Step 2A – Prong Two: Practical Application
Under Step 2A Prong Two, it is determined whether the additional elements integrate the judicial exception into a practical application.
For claims 1–5 and 7–10, the additional elements consist of applying the abstract calculations using generic machine learning models and conventional data processing steps. The claims do not recite any improvement to battery hardware, sensing technology, or computing architecture. Instead, they merely apply abstract mathematical processing to battery data.
Such limitations represent insignificant extra-solution activity, including routine data gathering and analysis, and therefore do not integrate the abstract idea into a practical application.
Accordingly, claims 1–5 and 7–10 require further analysis under Step 2B.
Step 2B – Inventive Concept
The claims do not include additional elements sufficient to amount to “significantly more” than the judicial exception.
The use of hybrid battery models combining physics-based modeling with machine learning is well-understood, routine, and conventional, as evidenced by U.S. Patent No. 11,527,786 B1, which discloses:
receiving battery pack data including current, voltage, temperature, simulation data, manufacturer data, and fleet history
predicting battery remaining useful life using a physics-based model combined with a machine learning model
generating battery properties and transmitting predicted results
The ordered combination of steps in claims 1–5 and 7–10 amounts to no more than applying abstract mathematical evaluations and predictive modeling using conventional machine learning techniques to battery data, which is a known design approach. The claims do not improve the functioning of a computer, neural network, or battery system itself, but merely automate known analytical techniques.
Therefore, claims 1–5 and 7–10 do not provide an inventive concept sufficient to transform the abstract idea into a patent-eligible application.
Dependent Claims (Claims 2–5 and 7–10)
Claim 2 defines evaluating prediction accuracy against a testing dataset and continuing training until a threshold is met. This recites iterative accuracy evaluation and optimization, which is a mathematical performance assessment.
Claim 3 defines calculating RMSE and refining LSTM models based on the RMSE. This recites mathematical error calculation and parameter adjustment.
Claim 4 limits validation to middle and late cycle life periods. This represents analytical selection of historical data.
Claim 5 limits validation to a constant 1C charge or discharge rate. This constrains operating conditions for analysis without technological improvement.
Claim 7 defines determining delay time and embedding dimension for reconstruction. This recites numerical parameter selection for mathematical transformation.
Claim 8 defines normalizing time-series data prior to reconstruction. This recites mathematical data scaling.
Claim 9 defines reducing noise through delayed phase state space processing. This recites mathematical filtering and transformation.
Claim 10 defines when SOC (430) and SOH (440) estimations are performed. This specifies timing of abstract calculations.
Each of these limitations remains within the judicial exception and does not add significantly more.
Claim Found Patent-Eligible under 35 U.S.C. § 101
Claim 6 recites that the time-series data are collected from a battery management system (BMS) integrated with the battery pack, wherein the BMS is configured to record operational parameters at predetermined intervals during charging and discharging cycles of the battery pack.
This limitation meaningfully integrates any underlying data analysis into a practical application by tying the claimed method to a real-world battery management system performing physical monitoring of battery operation. The claim is not directed merely to abstract data processing but instead specifies how and where the data is obtained, namely through a BMS that interacts with the battery pack during actual charging and discharging cycles.
The recited BMS-based data collection imposes a technological constraint on the claimed method and reflects an improvement to battery monitoring and management systems, as it enables structured acquisition of operational parameters in a real operating environment. Accordingly, Claim 6 integrates any judicial exception into a practical application and is therefore eligible under 35 U.S.C. § 101.
Claim Rejections - 35 USC § 103
The following is a quotation of 35 U.S.C. 103 which forms the basis for all obviousness rejections set forth in this Office action:
A patent for a claimed invention may not be obtained, notwithstanding that the claimed invention is not identically disclosed as set forth in section 102, if the differences between the claimed invention and the prior art are such that the claimed invention as a whole would have been obvious before the effective filing date of the claimed invention to a person having ordinary skill in the art to which the claimed invention pertains. Patentability shall not be negated by the manner in which the invention was made.
Claim(s) 1-19 is/are rejected under 35 U.S.C. 103 as being unpatentable over Park et al. (U.S. 2024/0110984 A1) in view of Chamali et al. (U.S. 2020/0081070 A1).
Regarding claim 1, Regarding claim 1, Park et al. disclose a machine learning method of estimating State of Charge (SOC 430) and State of Health (SOH 440) of a battery pack, comprising: collecting a plurality of time-series data related to operational parameters of a battery pack (battery measurements including voltage, current, temperature collected via BMS, paragraphs [0040]); processing said collected time-series data utilizing a nonlinear transformation and feature construction to represent battery state evolution (model fusion, time-series based ML feature construction, paragraphs [0037]); training a machine learning model using reconstructed representations of battery behavior for predicting SOH (440) of said battery pack (ML-based SOH (440) models trained on time-series data, paragraphs [0056]); feeding said time-series data to the trained model to predict an estimated SOH (440) value (SOH (440) prediction outputs, paragraph [0120]); training a machine learning model for predicting SOC (430) of the battery pack using time-series representations and model outputs (ML-based SOC (430) models including LSTM networks, paragraphs [0052-0053]); and feeding time-series data to obtain an estimated SOC (430) value (SOC (430) prediction in real-time, paragraphs [0053 & 0060]).
Park et al. do not explicitly disclose processing the time-series data utilizing a Nonlinear State Space Reconstruction (NSSR) algorithm to reconstruct first and second phase state space reconstructions, nor explicitly disclose separate phase reconstructions taking into account an estimated SOH (440) value.
Chamali et al. disclose processing the time-series data utilizing a Nonlinear State Space Reconstruction (NSSR) algorithm to reconstruct first and second phase state space reconstructions, nor explicitly disclose separate phase reconstructions taking into account an estimated SOH (440) value (see [0060] wherein processing time-series battery measurements using recurrent neural networks that reconstruct internal battery state trajectories through nonlinear temporal state evolution; RNN and LSTM hidden state evolution accumulating historical measurements, paragraph [0067]; reconstructing state representations from delayed and historical measurements; use of sequences of voltage, current, temperature to form nonlinear state representations, paragraphs [0057], [0069], [0096] and adapting model behavior based on changing conditions such as temperature and cycle variation; model adaptation and nonlinear state evolution, paragraphs [0067], [0096] and using reconstructed internal states as inputs to subsequent estimation tasks; SOC and SOH (440) estimation using learned state vectors, paragraphs [0066], [0146], [0147]).
Therefore, it would have been obvious to one of ordinary skill in the art, prior to the effective filing date, to modify Park et al. by incorporating nonlinear state space reconstruction as taught by Chamali et al., as doing so would enable explicit reconstruction of battery phase state spaces from time-series data, improve robustness to noise and hysteresis, and allow sequential SOC and SOH estimation using reconstructed internal state representations, because Chamali et al. emphasize that nonlinear recurrent state reconstruction from delayed time-series data improves estimation accuracy under varying operating conditions (see Chamali’s paragraphs [0067] & [0096]).
PNG
media_image1.png
1108
1506
media_image1.png
Greyscale
Regarding claim 2, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose evaluating accuracy of estimated SOH (440) and SOC (430) values using testing datasets and continuing training until performance metrics reach target thresholds (model validation using test datasets and performance metrics such as MAE and RMSE, paragraphs [0115]).
Regarding claim 3, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose validating trained models using testing datasets, calculating root mean square error (RMSE) between estimated and measured SOC (430) and SOH (440) values, and refining models based on RMSE (validation using RMSE and iterative refinement, paragraphs [0115]).
Regarding claim 4, Park et al. & Chamali et al. disclose the method of claim 3, wherein Park et al. further disclose validating SOC (430) and SOH (440) models over middle and late portions of battery life cycles through long-term cycling datasets and aging evaluations (cycle life evaluation and validation across aging stages, paragraph [0121]).
Regarding claim 5, Park et al. & Chamali et al. disclose the method of claim 3, wherein Park et al. further disclose validation using standardized constant-rate charge and discharge profiles including constant C-rate cycling (use of standardized drive cycles and controlled charge-discharge rates, paragraph [0090]).
Regarding claim 6, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose collecting time-series data from a battery management system integrated with a battery pack, configured to record operational parameters during charging and discharging cycles (BMS data acquisition, paragraph [0105]).
Regarding claim 7, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose constructing time-dependent representations of battery behavior from time-series data but do not explicitly disclose determining delay time and embedding dimension (via feature-level temporal modeling, paragraphs [0038] & [0056]); Chamali et al. disclose reconstructing battery state representations using delayed measurements and temporal embedding inherent in recurrent neural network architectures (use of historical sequences and delayed inputs forming embedding dimensions, paragraphs [0115]).
Regarding claim 8, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose normalizing time-series battery data prior to model processing to improve training stability and accuracy (data normalization and preprocessing, paragraphs [0056], [0090]).
Regarding claim 9, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose reducing noise effects through model fusion and temporal smoothing (noise mitigation via model fusion, paragraphs [0110]), and Chamali et al. disclose reducing instantaneous noise interference through delayed temporal state reconstruction inherent in LSTM state evolution (delayed state representation reducing noise sensitivity, paragraph [0067]).
Regarding claim 10, Park et al. & Chamali et al. disclose the method of claim 1, wherein Park et al. further disclose estimating SOC (430) continuously in real time during battery operation and estimating SOH (440) periodically at defined operational points such as cycle boundaries (real-time SOC (430) and cycle-based SOH (440) estimation, paragraphs [0059]).
Regarding claim 11, Park et al. disclose a system comprising a data acquisition module configured to collect time-series battery operational data (BMS data acquisition, paragraph [0033]); a data processing module incorporating nonlinear modeling to reconstruct time-series representations for SOC (430) and SOH (440) estimation (model fusion and temporal modeling, paragraph [0042]); an estimation module comprising machine learning models for SOC (430) and SOH (440) estimation including LSTM networks (ML SOC (430) and SOH (440) models, paragraphs [0117]); and a training module configured to train models until accuracy thresholds are met (iterative training with validation, paragraphs [0090] & [0115]).
Park et al. do not explicitly disclose neural network model is configured to be trained using the first reconstruct phase state space as input to predict an estimated SOH (440) value.
Chamali et al. disclose neural network model is configured to be trained using the first reconstruct phase state space as input to predict an estimated SOH value (see [0067], wherein reconstructing nonlinear battery state representations through recurrent neural network hidden states and temporal embeddings, nonlinear state reconstruction via LSTM hidden states, paragraph [0059] & [0096).
Therefore, it would have been obvious to one of ordinary skill in the art, prior to the effective filing date, to modify Park et al. by incorporating nonlinear state space reconstruction as taught by Chamali et al., as doing so would enable explicit reconstruction of battery phase state spaces from time-series data, improve robustness to noise and hysteresis, and allow sequential SOC and SOH estimation using reconstructed internal state representations, because Chamali et al. emphasize that nonlinear recurrent state reconstruction from delayed time-series data improves estimation accuracy under varying operating conditions (see Chamali’s paragraphs [0067] & [0096]).
Regarding claim 12, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose validation modules calculating RMSE between estimated and measured SOC (430) and SOH (440) values and refining models accordingly (validation and refinement, paragraphs [0115], [0121]).
Regarding claim 13, Park et al. & Chamali et al. disclose the system according to claim 12, wherein Park et al. further disclose validation using controlled constant-rate charge-discharge conditions (standardized cycling profiles, paragraphs [0090], [0121]).
Regarding claim 14, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose temporal modeling of time-series battery data (paragraph [0056]), and Chamali et al. disclose determining effective delay and embedding dimensions through temporal sequence modeling inherent in recurrent neural networks (paragraphs [0056] & [0119]).
Regarding claim 15, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose normalization and preprocessing of time-series data to mitigate outliers and improve model robustness (data preprocessing and normalization, paragraphs [0059]).
Regarding claim 16, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose reducing noise through temporal smoothing and model fusion (see paragraphs [0038], [0052]), and Chamali et al. disclose delayed temporal state reconstruction reducing instantaneous noise effects (paragraphs [0056] & [0119]).
Regarding claim 17, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose real-time SOC (430) estimation and periodic SOH (440) estimation at defined operational points (see paragraph [0036]).
Regarding claim 18, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose adjusting model parameters based on environmental factors including temperature (temperature-aware model training and adjustment, paragraphs [0056], [0078]).
Regarding claim 19, Park et al. & Chamali et al. disclose the system according to claim 11, wherein Park et al. further disclose using state vectors comprising SOC (430) and SOH (440) values as inputs to machine learning models for training and estimation (see state vectors and model inputs, paragraphs [0037] & [0108]).
Conclusion
The prior art made of record and not relied upon is considered pertinent to applicant's disclosure.
U.S. 11,065,978 B2 to Aykol et al. disclose Systems, methods, and storage media for optimizing performance of a vehicle battery pack are disclosed. A method includes receiving data pertaining to cells within a battery pack installed in each vehicle of a fleet of vehicles, the data received from at least one of each vehicle, providing the data to a machine learning server, and directing the machine learning server to generate a predictive model. The predictive model is based on machine learning of the data. The method further includes providing the predictive model to each vehicle, the predictive model providing instructions for adjusting configuration parameters for each of the cells in the battery pack such that the battery pack is optimized for a particular use, and directing each vehicle to optimize performance of the vehicle battery pack based on the predictive model.
U.S. 2021/0173012 A1 to Subbotin et al. disclose a battery management system includes a memory, a current sensor that measures a current flow through a battery to a load, a voltage sensor that measures a voltage level between a first terminal and a second terminal of the battery that are each connected to the load, and the memory, a temperature sensor that measures a temperature level of the battery; and a controller configured to be operatively connected to the current sensor, temperature sensor, and voltage sensor. The controller is configured to receive a measurement of a first current level and a first voltage level and utilize a corrected capacity and corrected open circuit voltage estimate to output an estimated open circuit voltage of the battery as compared to an estimated capacity.
U.S. 2020/0271725 A1 to Herring et al. disclose systems, methods, and storage media for generating a predicted discharge profile of a vehicle battery pack are disclosed. A method includes receiving, by a processing device, data pertaining to cells within a battery pack installed in each vehicle of a fleet of vehicles operating under a plurality of conditions, the data received from at least one of each vehicle in the fleet of vehicles, providing, by the processing device, the data to a machine learning server, directing, by the processing device, the machine learning server to generate a predictive model, the predictive model based on machine learning of the data, generating, by the processing device, the predicted discharge profile of the vehicle battery pack from the predictive model, and providing the discharge profile to an external device.
Any inquiry concerning this communication or earlier communications from the examiner should be directed to TRUNG NGUYEN whose telephone number is (571)272-1966. The examiner can normally be reached on Mon- Friday 8AM - 4:00PM Eastern Time. If attempts to reach the examiner by telephone are unsuccessful, the examiner’s supervisor, Huy Phan can be reached on 571-272-7924. The fax phone number for the organization where this application or proceeding is assigned is 571-273-8300. Information regarding the status of an application may be obtained from the Patent Application Information Retrieval (PAIR) system. Status information for published applications may be obtained from either Private PAIR or Public PAIR. Status information for unpublished applications is available through Private PAIR only. For more information about the PAIR system, see http://pair-direct.uspto.gov. Should you have questions on access to the Private PAIR system, contact the Electronic Business Center (EBC) at 866-217-9197 (toll-free). If you would like assistance from a USPTO Customer Service Representative or access to the automated information system, call 800-786-9199 (IN USA OR CANADA) or 571-272-1000.
Examiner: /Trung Q. Nguyen/- Art 2858
December 17, 2025
/HUY Q PHAN/ Supervisory Patent Examiner, Art Unit 2858
Prosecution Insights