A Review: Research Progress in Bridge Structural Health Monitoring From the Perspective of AI Development

Bridge structural health monitoring (BSHM) has consistently been a research hotspot in civil engineering. The field of BSHM has experienced a significant transition from traditional manual inspections to an advanced integration of artificial intelligence (AI), culminating in the current peak with data‐driven AI methodologies. Nevertheless, despite the impressive performance, data‐driven AI techniques such as machine learning (ML) and DL exhibit limitations in interpretability, stability, and security. Conversely, the earlier generation of knowledge‐driven AI, including expert systems and logical reasoning, while offering greater interpretability and stability, has not achieved widespread adoption due to its limited scope, inefficiency, and subpar predictive accuracy. Against this backdrop, the current paper advocates for the creation of more reliable and intelligible explainable artificial intelligence (XAI). The paper provides a chronological overview of AI’s evolution within BSHM and discusses the fundamental principles of knowledge‐driven AI, data‐driven AI, and XAI. It examines their respective applications in BSHM and evaluates the advantages and limitations of these approaches. The paper concludes by anticipating future trends and identifying the challenges within the field. The findings underline the necessity for advancement in XAI in BSHM. The envisioned AI is designed to incorporate the advantages of both traditional knowledge‐driven AI and data‐driven AI while minimizing their respective shortcomings. This symbiosis is projected to set the direction for AI’s progression in BSHM.