Andi Prasetiawan, Mochamad Subchan Mauludin, Singgih Dwi Prasetyo, Yuki Trisnoaji, Darul Prayogo
Offshore wind turbine blades operate under extreme marine stressors that accelerate structural degradation and increase the complexity of maintenance. Recent advances in deep reinforcement learning (DRL) and haptic-feedback robotics offer opportunities to enhance precision, safety, and automation in composite blade repair. This study presents a systematic review and meta-analysis that synthesizes empirical evidence on DRL–haptic robotic systems for offshore blade maintenance—a PRISMA-guided, multi-database search identified 20 eligible studies published between 2019 and 2025. Data extraction incorporated quantitative metrics on task accuracy, operational efficiency, and PHM performance, which were synthesized using random-effects modeling. The pooled effect size (SMD = 0.800; 95% CI: 0.740–0.870) demonstrates significant improvements over conventional methods, supported by moderate heterogeneity (I² = 45.2%). Subgroup analyses reveal that PPO-based DRL models and force-feedback haptics deliver the most substantial performance gains. Despite promising results, limitations persist in short-duration testing, laboratory-focused validation, and inconsistent evaluation standards. Overall, the evidence indicates that DRL-enabled haptic robotics is a maturing technology with substantial potential to enhance offshore blade repair reliability, reduce human risk exposure, and advance next-generation PHM strategies. © 2026 The authors. This article is published by IIETA and is licensed under the CC BY 4.0 license (http://creativecommons.org/licenses/by/4.0/).
Politeknik Ilmu Pelayaran Semarang, Merchant Marine Polytechnic Semarang, Semarang, 50242, Indonesia; Department of Informatics Engineering, Universitas Wahid Hasyim, Semarang, 50224, Indonesia; Power Plant Engineering Technology, State University of Malang, Malang, 65145, Indonesia
