Nazrul Amin

Nazrul Amin

Auto-generated profile, based on publicly available information

Nazrul Amin

Department of Computer Engineering, Jeju National University, Jeju-si, South Korea

Research Focus & Metrics

Academic Career Length

2Years (2025-Present)

(Calculated once Scilit publication data is synced: current year − year of first publication + 1)

Representative Publications

View All Publications
12 September 2025 in Engineering Reports
Verified
Journal Article
Open access
Imrus SalehinImrus Salehin
Imrus Salehin
TSTomal Ahmed Sajib
Tomal Ahmed Sajib
Nazmul Huda BadhonNazmul Huda Badhon
Nazmul Huda Badhon
SRSakibul Hassan Rifat
Sakibul Hassan Rifat
NANazrul Amin
Nazrul Amin
NMNazmun Nessa Moon
Nazmun Nessa Moon
Large language models (LLMs), like the GPT series, have recently emerged as transformative tools in the medical field due to their human-like language generation and understanding. This systematic review examines the evolution, applications, and challenges of medical LLMs in digital health and clinical technology. A structured search was conducted across ScienceDirect, PubMed, Scopus, and manual sources from 2007 to 2025, following PRISMA 2020 guidelines. After applying inclusion and exclusion criteria, 185 studies were selected from an initial pool of 698 papers. Among these, 30 representative studies were analyzed in-depth based on their relevance, methodological quality, and contribution to diverse LLM applications in health care. Most research centered on GPT-based models, with over 81% demonstrating strong performance in language generation, diagnostic assistance, and clinical documentation, based on automated metrics and human feedback. Notably, some models achieved up to 90% satisfaction from healthcare professionals. The findings reveal LLM's potential to enhance patient interaction, decision support, and overall healthcare efficiency. This review contributes by synthesizing key advancements, assessing model performance, and outlining ethical challenges such as trust, privacy, and safe deployment. It offers novel insights for researchers and practitioners seeking to adopt or improve LLM integration in health care. Future directions include improving transparency, developing domain-specific models, and establishing regulatory frameworks for responsible use.

6 January 2026 in Energy Reports
Verified
Journal Article
Open access
NANazrul Amin
Nazrul Amin
YKYong-Woon Kim
Yong-Woon Kim
CKChulung Kang
Chulung Kang
YBYung-Cheol Byun
Yung-Cheol Byun

9 June 2026 in Computers and Electrical Engineering
Verified
Journal Article
Closed access
NANazrul Amin
Nazrul Amin
YKYong Woon Kim
Yong Woon Kim
Yung-Cheol ByunYung-Cheol Byun
Yung-Cheol Byun

3 November 2025 in IEEE Access
Verified
Journal Article
Open access
NANazrul Amin
Nazrul Amin
YKYong Woon Kim
Yong Woon Kim
Yung-Cheol ByunYung-Cheol Byun
Yung-Cheol Byun
Automated Guided Vehicles (AGVs) are central to material handling in modern warehouses, where precise motion control is critical for ensuring high throughput, safety, and energy efficiency. At the core of this control lies Proportional-Integral-Derivative (PID) tuning, which remains a persistent challenge in dynamic, data-scarce, and resource-constrained environments. This study introduces a prompt-driven PID tuning framework that leverages large language models (LLMs) to infer optimal gains through structured in-context learning, eliminating the need for system identification or iterative tuning. The proposed framework implements two novel strategies: (1) MultiSpeed (MS) prompting, which interpolates gain values based solely on the target speed, and (2) Few-Shot + MultiSpeed (FS+MS) prompting, which incorporates exemplar-based conditioning for improved generalization. Experiments conducted on a two-wheeled AGV across five discrete velocities (5–25 m/min) demonstrate that both methods achieve PID gains closely aligned with simulation-optimized baselines, while FS+MS consistently yields the most accurate and stable results. The inferred gains preserve physically consistent trends, such as increasing Kp and decreasing Ki with speed, and maintain trajectory tracking RMSE within 5–10% of optimal values. Our framework achieves over 80% reduction in calibration effort and supports millisecond-scale inference on embedded hardware, making it well-suited for real-time deployment in edge robotics.