Ali Serdar Atalay (AI4SEC OÜ, Research and Development)
Edge-Cloud Digital Twins for Smarter EV Maintenance
The transition to electrified vehicles (EVs) is reshaping the global automotive landscape with a focus on efficiency, sustainability, and innovation. However, ensuring the reliability and safety of these advanced systems requires predictive maintenance (PdM) strategies that go beyond conventional approaches. Ali Serdar Atalay, during his secondment at ERARGE, collaborated with researchers to design cutting-edge edge-cloud digital twin frameworks to address these challenges.
This initiative aims to harness data-driven AI techniques, supported by an advanced semantic backend, to revolutionize PdM in the electrified automotive domain. The secondment, conducted from May 1st to August 14th, 2024, provided an opportunity to synergize efforts between AI4SEC and ERARGE to develop a trustworthy and intelligent framework for EV powertrains.
From Data Cleaning to Semantic Ontologies
One of the project’s core objectives was to review the state-of-the-art in predictive maintenance algorithms and edge-cloud frameworks, focusing on secure and reliable EV operations. The framework integrates batch and streaming data pre-processing from edge to cloud, enabling seamless data cleaning, harmonization, and feature engineering.
Through collaborative research, we developed a PdM-oriented ontology combining the Smart Sensor Network Ontology, ENISA’s cyber threat taxonomy, and a customized event ontology. This semantic backend facilitates:
- Enhanced Data Governance: Handling multidimensional, heterogeneous in-vehicle and out-vehicle data.
- Smarter Digital Twins: Enabling real-time insights and actionable predictions for EV systems.
- Advanced Cybersecurity: Incorporating hardware-based security features for resilient operations, Ali learned about Erarge Hardware Security Module (HSM) and worked with Ibrahim to connect to AI4SEC cloud.
Expanding Knowledge Through Collaboration and Research
The secondment emphasized cooperation with Turkish research organizations, including engagements with Zonguldak Bulent Ecevit University and OTOKAR Turkiye. The findings were presented at the KISES 2024 Symposium, focusing on:
- AI-powered Trustworthy Predictive Maintenance Solutions for EV Powertrains
- Driving Toward Sustainability: Technology and Industry Initiatives in EV Transition
The project also leveraged discussions with Tallinn Technical University and other international institutions to enhance the semantic framework’s robustness. A hybrid workshop conducted on July 9th, 2024, further disseminated insights to the global research community.
Bridging Academia and Industry
This research is set to propel the TEAMING consortium forward, aligning with the goals of sustainable, secure, and efficient EV systems. The comprehensive framework developed during the secondment showcases the potential of integrating physics-informed machine learning with semantic AI for trustworthy PdM solutions. Future work will focus on expanding the backend’s capabilities with IoT-enabled situational awareness, explainable AI, and cyber-resilient cloud architectures to maximize the potential of connected and autonomous EVs.
LinkedIn Abstract:
.Ali Serdar Atalay, a researcher at AI4SEC, recently completed a secondment at ERARGE, where he collaborated on developing cutting-edge semantic AI frameworks for predictive maintenance (PdM) in the electrified automotive domain. During his time at ERARGE from May to August 2024, Atalay worked to design an advanced edge-cloud digital twin architecture that leverages data-driven AI techniques and a robust semantic backend. The goal was to revolutionize PdM for electric vehicle (EV) powertrains, enhancing reliability, safety, and sustainability.
Key aspects of the framework include:
- Integrating batch and streaming data preprocessing capabilities for seamless data cleaning, harmonization, and feature engineering.
- Developing a PdM-focused ontology that enables advanced data governance, smarter digital twins, and improved cybersecurity.
- Harnessing hardware-based security features for resilient EV operations.
Atalay’s work involved close collaboration with Turkish research organizations like ERARGE Turkiye. The findings were presented at the KISES 2024 Symposium, covering AI-powered predictive maintenance solutions and technology/industry initiatives in the EV transition.
This research aligns with the goals of the TEAMING consortium, showcasing the potential of integrating physics-informed machine learning and semantic AI for trustworthy PdM in connected and autonomous electric vehicles. Future work will focus on expanding the framework’s capabilities in areas like IoT-enabled situational awareness, explainable AI, and cyber-resilient cloud architectures.
Event: Participated and discussed Teaming project with Barcelona Super Computing, ODTU and Tubitak Turkey at Conference Basarım in May 2024
https://indico.truba.gov.tr/event/140
Alper Kanak, Serhat Ege İnanç, Sercan Tanrıseven, İbrahim Arif, Cengiz Bektaş, Oguzhan Herkiloğlu, Ali Serdar Atalay, Salih Ergün, “What About the Energy-Efficiency of Complementary Services making a Fuel Cell Electrical Vehicle more Trustworthy and AI-Powered?”, submitted to IEEE Elmar Conference (invited paper), 2024.
https://ieeexplore.ieee.org/abstract/document/10694222
We published two conference and presented for Karaelmas International Science and Engineering Symposium (KISES 2024) Zonguldak, Turkiye, 10-11 May 2024 with large bus and truck manufacturer OTOKAR Turkiye.
An AI-powered Trustworthy Predictive Maintenance Solution for Electrical Vehicle Powertrain using Physics-Informed Machine Learning
Driving Toward a Sustainable Future: Technology, Economics, and Industry Initiatives in the Transition to Connected and Autonomous Electric Vehicles
https://kises.beun.edu.tr/duyurular/abstract-book-has-been-published..html