Review on fault diagnosis and fault-tolerant control schemes for robotic manipulators: Recent advances in artificial intelligence and digital twin technologies

Fault-tolerant control (FTC) is essential for ensuring the safety, reliability, and operational continuity of robotic manipulators in modern industrial and safety-critical applications. This review presents a comprehensive and systematic survey of FTC schemes for robotic manipulators, tracing their evolution from classical model- and signal-based approaches to contemporary data-driven and intelligent frameworks. The study is conducted in two phases: first, a historical analysis (1975–2000) examining foundational developments in hardware redundancy, analytical redundancy, observer-based techniques, and robust/adaptive control; second, a detailed review (2000–2025) highlighting recent advances enabled by artificial intelligence (AI), machine learning (ML), deep learning (DL), reinforcement learning (RL), and digital twin technologies (DTT). The review categorizes existing literature along three main axes: (i) the progression of model-based and signal-based FTC strategies, (ii) the evolving role of sensors and sensor-fusion techniques in fault diagnosis and reconfiguration, and (iii) the integration of artificial intelligence techniques, including machine learning for fault diagnosis, predictive maintenance, real-time monitoring, and autonomous fault recovery. A comparative analysis is provided to clarify the strengths and limitations of model-based, data-driven, and hybrid FTC approaches. By bridging classical control theory with intelligent and cyber–physical paradigms, this work highlights a paradigm shift toward proactive, adaptive, and self-learning fault-tolerant systems. The review offers structured insights, identifies current challenges, and outlines future research directions to support the development of resilient robotic manipulators across manufacturing, healthcare, aerospace, logistics, and agriculture applications.