Industrial robot fault detection and diagnosis for intelligent maintenance: A review of data, methods, and emerging trends

Abstract

Industrial robots are widely used in intelligent manufacturing due to their high efficiency, flexibility, and repeatability. However, long-term operation under complex working conditions inevitably leads to performance degradation and various faults across mechanical, transmission, and electrical subsystems. Owing to the structural complexity of industrial robots, strong coupling among subsystems, and heterogeneous monitoring data, fault detection and diagnosis has become a challenging yet crucial task for ensuring system reliability and enabling intelligent maintenance. Despite considerable progress, existing studies remain fragmented across different robot subsystems, data modalities, and diagnostic paradigms, and a systematic review dedicated to industrial robot fault detection and diagnosis is still lacking. To address this gap, this paper presents a comprehensive and structured review from three complementary perspectives: fault characteristics and data resources, diagnostic methodologies, and emerging trends. Specifically, the review summarizes typical fault types and data acquisition strategies, organizes existing diagnostic methods into a structured taxonomy, and discusses key challenges and emerging directions such as multimodal fusion, trustworthy deployment, and digital-twin-driven diagnosis. This review provides a structured reference for advancing both theoretical research and engineering applications of industrial robot fault detection and diagnosis in intelligent manufacturing systems.