What is KTBIX?​

KTBIX brings together two world-class institutions — Sejong University (Seoul, Korea) and Istanbul Technical University (Istanbul, Türkiye) — to tackle one of the most pressing challenges in electric vehicle technology: accurate, robust battery State-of-Health (SOH) estimation under real-world field conditions.

Our core technical innovation is the PI-VIB-ResNet architecture — a Physics-Informed Variational Information Bottleneck combined with a Residual Neural Network — that directly addresses the limitations of purely data-driven EIS-based battery diagnostics.

Building on Published Foundations​

KTBIX directly extends our 2025 joint publication in Scientific Reports (DOI: 10.1038/s41598-025-12771-4), where we established the AE-BPNN baseline for battery SOH estimation via EIS. The PI-VIB-ResNet architecture is expected to improve upon this baseline by ≥15% in RMSE (target: ≤1.2%).

What's Coming​

Over the next 24 months, you can expect:

• 📡 WP1 — A new open industrial EIS dataset from Turkish T-EV platforms and TOGG vehicles
• 🧠 WP2 — Open-source PI-VIB-ResNet implementation (PyTorch)
• 📊 WP3 — Rigorous benchmark comparisons against 5+ state-of-the-art methods
• 🌐 WP4 — Two international workshops (Istanbul 2027, Seoul 2028)
• 📄 2+ Q1 journal publications

Stay Connected​

Follow the project at ktbix.org for news, dataset releases, and workshop announcements.

— The KTBIX Team (SJU · ITU)