SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models

The state of health (SOH) of a Li-ion battery is determined by complex interactions among its internal components and external factors. Approaches leveraging deep learning architectures have been proposed to predict the SOH using convolutional networks, recurrent networks, and transformers. Recently...

Full description

Saved in:
Bibliographic Details
Main Authors: Jose Ignacio Olalde-Verano, Sascha Kirch, Clara Perez-Molina, Sergio Martin
Format: Article
Language:English
Published: IEEE 2025-01-01
Series:IEEE Access
Subjects:
Online Access:https://ieeexplore.ieee.org/document/10818656/
Tags: Add Tag
No Tags, Be the first to tag this record!
_version_ 1850085344447299584
author Jose Ignacio Olalde-Verano
Sascha Kirch
Clara Perez-Molina
Sergio Martin
author_facet Jose Ignacio Olalde-Verano
Sascha Kirch
Clara Perez-Molina
Sergio Martin
author_sort Jose Ignacio Olalde-Verano
collection DOAJ
description The state of health (SOH) of a Li-ion battery is determined by complex interactions among its internal components and external factors. Approaches leveraging deep learning architectures have been proposed to predict the SOH using convolutional networks, recurrent networks, and transformers. Recently, Mamba selective state space models have emerged as a new sequence model that combines fast parallel training with data efficiency and fast sampling. In this paper, we propose SambaMixer, a Mamba-based model for predicting the SOH of Li-ion batteries using multivariate time signals measured during the battery’s discharge cycle. Our model is designed to handle analog signals with irregular sampling rates and recuperation effects of Li-ion batteries. We introduce a novel anchor-based resampling method as an augmentation technique. Additionally, we improve performance and learn recuperation effects by conditioning the prediction on the sample time and cycle time difference using positional encodings. We evaluate our model on the NASA battery discharge dataset, reporting MAE, RMSE, and MAPE. Our model outperforms previous methods based on CNNs and recurrent networks, reducing MAE by 52%, RMSE by 43%, and MAPE by 7%.
format Article
id doaj-art-c1f2890fdb9f4deeb5eb3fb433112c7e
institution DOAJ
issn 2169-3536
language English
publishDate 2025-01-01
publisher IEEE
record_format Article
series IEEE Access
spelling doaj-art-c1f2890fdb9f4deeb5eb3fb433112c7e2025-08-20T02:43:43ZengIEEEIEEE Access2169-35362025-01-01132313232710.1109/ACCESS.2024.352432110818656SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space ModelsJose Ignacio Olalde-Verano0https://orcid.org/0000-0001-8058-156XSascha Kirch1https://orcid.org/0000-0002-5578-7555Clara Perez-Molina2https://orcid.org/0000-0001-8260-4155Sergio Martin3https://orcid.org/0000-0002-4118-0234Department of Electric and Computer Engineering, UNED-Universidad Nacional de Educación a Distancia, Madrid, SpainDepartment of Electric and Computer Engineering, UNED-Universidad Nacional de Educación a Distancia, Madrid, SpainDepartment of Electric and Computer Engineering, UNED-Universidad Nacional de Educación a Distancia, Madrid, SpainDepartment of Electric and Computer Engineering, UNED-Universidad Nacional de Educación a Distancia, Madrid, SpainThe state of health (SOH) of a Li-ion battery is determined by complex interactions among its internal components and external factors. Approaches leveraging deep learning architectures have been proposed to predict the SOH using convolutional networks, recurrent networks, and transformers. Recently, Mamba selective state space models have emerged as a new sequence model that combines fast parallel training with data efficiency and fast sampling. In this paper, we propose SambaMixer, a Mamba-based model for predicting the SOH of Li-ion batteries using multivariate time signals measured during the battery’s discharge cycle. Our model is designed to handle analog signals with irregular sampling rates and recuperation effects of Li-ion batteries. We introduce a novel anchor-based resampling method as an augmentation technique. Additionally, we improve performance and learn recuperation effects by conditioning the prediction on the sample time and cycle time difference using positional encodings. We evaluate our model on the NASA battery discharge dataset, reporting MAE, RMSE, and MAPE. Our model outperforms previous methods based on CNNs and recurrent networks, reducing MAE by 52%, RMSE by 43%, and MAPE by 7%.https://ieeexplore.ieee.org/document/10818656/Li-ion batterymambastate space modelstate of health predictionmultivariate time seriesdeep learning
spellingShingle Jose Ignacio Olalde-Verano
Sascha Kirch
Clara Perez-Molina
Sergio Martin
SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
IEEE Access
Li-ion battery
mamba
state space model
state of health prediction
multivariate time series
deep learning
title SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
title_full SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
title_fullStr SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
title_full_unstemmed SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
title_short SambaMixer: State of Health Prediction of Li-Ion Batteries Using Mamba State Space Models
title_sort sambamixer state of health prediction of li ion batteries using mamba state space models
topic Li-ion battery
mamba
state space model
state of health prediction
multivariate time series
deep learning
url https://ieeexplore.ieee.org/document/10818656/
work_keys_str_mv AT joseignacioolaldeverano sambamixerstateofhealthpredictionofliionbatteriesusingmambastatespacemodels
AT saschakirch sambamixerstateofhealthpredictionofliionbatteriesusingmambastatespacemodels
AT claraperezmolina sambamixerstateofhealthpredictionofliionbatteriesusingmambastatespacemodels
AT sergiomartin sambamixerstateofhealthpredictionofliionbatteriesusingmambastatespacemodels