Applied research
Experience carrying research from data preparation and model development through evaluation, publication, open datasets, and software.
Available from January 2027
Anomaly detection · Explainable diagnosis · Energy systems
Marwa Zitouni
My research concerns anomaly detection and early fault diagnosis in energy systems. I work with sensor time series, infrared thermal imaging, and the fusion of the two, combining machine learning with knowledge-based reasoning so that each detection is accompanied by an explanation on which an engineer can act.
5Published peer-reviewed papers
3Submitted and invited manuscripts
2Technical deliverables contributed within ENERGETIC
01/2027Available for a new research position
Professional profile
What I bring
I combine artificial intelligence, engineering knowledge, and scientific communication to develop diagnostic methods that are both accurate and understandable.
Engineering perspective
A computer-science research profile grounded in electrical engineering and focused on practical energy-system problems.
Collaborative delivery
Experience contributing to technical deliverables and publications within a European consortium of academic and industrial partners.
Working areas
Three connected areas of expertise
My work follows the path from multimodal data generation to anomaly detection and knowledge-supported diagnosis.
I build and curate synchronized battery datasets combining electrical sensor time series with infrared thermal images.
Sensor measurementsVoltage, current, temperature, and related operating variables.
Thermal imagingInfrared image sequences aligned with the electrical measurements.
I develop unsupervised and multimodal models that learn normal operating behaviour and identify abnormal patterns at an early stage.
Time-series modelsAutoencoders and recurrent models for sensor-based anomaly detection.
Multimodal modelsFusion of time-series and thermal-image information for stronger detection.
I connect detected anomalies with structured battery knowledge so that diagnostic results are traceable, interpretable, and useful for engineering decisions.
Knowledge representationOntologies and knowledge graphs for battery faults, sensors, and observations.
Reasoning and explanationSemantic rules and stream reasoning that link evidence to a diagnosis.
European project contribution
Horizon Europe ENERGETIC
Research conducted within an international consortium working on advanced sensing, battery monitoring, artificial intelligence, and explainable diagnosis.
From multimodal battery data to understandable diagnostics.
2023–2026 · ICube / INSA Strasbourg
Visit the ENERGETIC project ↗As a PhD researcher in ENERGETIC, I have contributed research and technical content across battery data generation, anomaly detection, and knowledge-based explainable diagnosis.
2Technical project deliverables contributed within the consortium.
3Connected research areas: multimodal data, detection, and explainable diagnosis.
EuropeanCollaboration with academic and industrial partners.
Research outputsPublished papers, submitted and invited manuscripts, datasets, software, and project reporting.
Career
Experience and education
My path connects artificial intelligence, electrical engineering, renewable energy, and international research collaboration.
Experience
PhD Researcher — ICube Laboratory, INSA Strasbourg
Research on artificial intelligence for lithium-ion battery monitoring within the Horizon Europe ENERGETIC project, including contributions to two technical deliverables.
Electrical Engineer — Elsewedy Electric, Algeria
Worked in an industrial engineering environment involving electrical systems, planning, and operational coordination.
Power Systems Engineering Intern — SKTM, Adrar
Gained practical exposure to wind-farm operation, monitoring, and grid-integration constraints.
Education
PhD in Computer Science / Artificial Intelligence
INSA Strasbourg — Université de Strasbourg. Thesis: Explainable Knowledge-Based Diagnosis for Lithium-Ion Batteries.
Engineering Degree + Master’s in Electrical Engineering
École Nationale Polytechnique d’Alger (ENP) — Preparatory engineering studies, 2015–2017.
ESGEE Oran — Electrical networks and power systems, 2017–2020.
SUSI Fellow — U.S. Department of State
Competitively selected academic leadership programme including academic residency, workshops, and community engagement in Massachusetts, United States.
Publications and manuscripts
Research contributions
Published work and submitted or invited manuscripts covering multimodal anomaly detection, semantic reasoning, knowledge representation, and battery state estimation.
Published peer-reviewed work
Five published contributions, including four first-author papers and one IEEE journal paper as co-author.
Toward Early Battery Anomaly Detection: A Multimodal Dataset and Deep Learning Framework for Lithium-Ion Batteries
Energy and AI, article 100816
Journal article
A Stream Reasoning Framework for Thermal Image-Based Anomaly Detection in Lithium-Ion Batteries
ICAART 2026, volume 2, pp. 1589–1600
Full conference paper
Spiking Neural Networks for Accurate and Efficient State of Health Estimation of Lithium-Ion Batteries Across Varying Temperatures
IEEE Open Journal of Vehicular Technology
Journal article · Co-authorCited by 1
Anomaly Detection in Lithium-Ion Batteries via Stream Reasoning on Structured Knowledge and Time-Series Data
Procedia Computer Science 270, pp. 3688–3697 · KES 2025 Best Paper Track
Conference paperCited by 2
Toward Anomaly Representation in Lithium-Ion Batteries: An Ontology-Based Approach
Procedia Computer Science 246, pp. 1319–1328
Conference paperCited by 4
Submitted and invited work
These manuscripts are separated from the published record and labelled by their current status.
Semantic Multi-Modal Framework for Early Stage Anomaly Detection in Lithium-Ion Batteries
Extended version of the KES 2025 paper submitted to Semantic Web Journal
Journal extension · Submitted
LLM-Assisted Ontology Enrichment for Explainable Lithium-Ion Battery Anomaly Diagnosis
Invited extension of the ICAART 2026 paper for a book chapter · Not yet published
Invited book chapter
BAO: A Modular Ontology for Li-Ion Battery Fault Diagnosis in Electric Vehicles
Submitted journal manuscript
Submitted
Citation counts reflect the Google Scholar values provided in August 2026.
Core capabilities
A focused research toolkit
Methods and tools used directly across my work on multimodal battery monitoring and explainable fault diagnosis.
Multimodal data
Preparation and alignment of heterogeneous sensor data.
Anomaly detection
Unsupervised and multimodal learning for early abnormal-behaviour detection.
Knowledge-based diagnosis
Structured knowledge and reasoning for traceable diagnostic conclusions.
Research software
A compact toolset supporting modelling, image processing, and reproducible research.
Application domainLithium-ion battery monitoring, anomaly detection, fault diagnosis, and safety-oriented decision support in energy systems.