Ph.D. Researcher · AI Mechatronics Laboratory (AML Lab)
Tech University of Korea · IT·Semiconductor Convergence Engineering
- 🔬 I develop AI-driven fault-diagnosis and predictive-maintenance systems
- 🤖 My research covers industrial robots, harmonic-drive reducers, wafer-transfer robots, and precision systems
- 🏭 I focus on reliable and explainable AI for semiconductor manufacturing equipment
- 📡 I work with acceleration, torque, encoder, vibration, current, temperature, and process data
- ⚙️ I integrate intelligent sensing with EtherCAT real-time control
- 🧩 My methods include signal processing, ML/DL, physics-informed learning, uncertainty quantification, and explainable AI
- 🌏 Originally from Bangladesh and currently based in South Korea
- 💬 Ask me about PHM, Industrial AI, Robotics, Physical AI, Agentic AI, RUL, and Smart Sensing
| Area | Focus |
|---|---|
| Prognostics & Health Management | Condition monitoring, fault detection, classification, and RUL |
| Industrial & Physical AI | Data-driven and physics-informed intelligence for manufacturing |
| Industrial Robotics | Harmonic-drive, robot-joint, path-accuracy, and vibration analysis |
| Semiconductor Equipment | Wafer-transfer robot reliability and belt/backlash diagnosis |
| Explainable & Uncertain AI | SHAP, LIME, Bayesian learning, Gaussian processes, and uncertainty |
| Agentic AI | Monitoring, diagnosis, decision, optimization, and compensation agents |
Data-Driven Approach for Fault Diagnosis of Harmonic Drives Using Wireless Acceleration Sensors and Machine Learning
Md. Saiful Islam, Kihyun Kim, and Hyo-Young Kim
International Journal of Precision Engineering and Manufacturing–Green Technology, 12, 951–968, 2025
99.81% Random Forest accuracy · Wireless sensing · SHAP and LIME
View publication
Md. Saiful Islam, Young Kyoung Ji, Kihyun Kim, and Hyo-Young Kim
International Journal of Precision Engineering and Manufacturing–Smart Technology, 2025
99.80% Random Forest accuracy · Semiconductor robotics · Predictive maintenance
View publication
➡️ See all publications on Google Scholar
Machine Learning: Random Forest · XGBoost · SVM · KNN · Logistic Regression · Naïve Bayes
Deep Learning: CNN · Transformer · ANN · Autoencoders · Bayesian Neural Networks
Signal Processing: FFT · STFT · CWT · Wavelets · Physics-based residuals
Control & Integration: EtherCAT · SOEM · CSP · HIL · Digital Twins
- Physics-informed diagnosis for mechatronic and optical systems
- Agentic AI for industrial monitoring, diagnosis, decision, and compensation
- Deep-learning diagnosis of harmonic-drive faults using torque and acceleration
- Predictive maintenance of wafer-transfer robot belt and backlash systems
- Aerostatic bearing optimization using Gaussian processes and PINNs
- HIL digital twins and real-time thermal-error compensation
Advancing reliable, explainable, and intelligent manufacturing systems.