Computational Intelligence and Automation for Resilient Bridge Infrastructure

• Reviews computational intelligence and automation technologies for bridge infrastructure. • Combines mixed scientometric and systematic methods on journal papers worldwide. • Maps trends in structural health monitoring, damage detection, predictive maintenance, and robotics. • Highlights the significant increase in research activity related to AI-enabled bridge operations. • Future needs include benchmarking, uncertainty, and recovery-linked decisions. Bridges are essential assets in transportation systems, and their failure or prolonged closure can disrupt emergency response, supply chains, and daily mobility. Aging bridge inventories face increasing demands from traffic growth, environmental exposure, and extreme events, making reliable in-service assessment and maintenance a structural engineering priority. This paper reviews how computational intelligence and automation are being used to support bridge condition evaluation and intervention decisions during the in-service phase by improving monitoring, inspection, damage assessment, and maintenance decision-making. A PRISMA-guided search was conducted in the Scopus database using title, abstract, and keyword terms related to bridges, computational intelligence, automation, and in-service operation. The initial search returned 109 records; after restricting the results to English-language journal articles published between 2000 and 2025 and screening for bridge-specific studies that applied computational intelligence and/or automation technologies to in-service monitoring, inspection, damage assessment, deterioration prediction, or maintenance decision-making, 37 studies were retained for combined scientometric and systematic analysis. The literature is organized into five application areas: structural health monitoring, damage detection and condition assessment, predictive maintenance and optimization, robotic and autonomous inspection, and post-event decision support. The review shows a clear rise in research activity since 2023, with deep learning leading advances in time-series monitoring and vision-based inspection, while unmanned and robotic platforms are making data collection safer and faster. However, most studies still rely on limited validation data, inconsistent performance reporting, and weak links between detected defects and decisions about service continuity, operability, and recovery. From a resilience perspective, these developments are most valuable when they reduce the time required to detect damage, classify safety and functionality, and prioritize interventions following disruptive events (e.g., floods, earthquakes, hurricanes, and climate-driven extremes). The review therefore interprets computational intelligence and automation not only as tools for routine condition assessment, but also as enablers of resilience-based management that supports continuity of transportation service, faster recovery, and risk-informed allocation of limited maintenance resources. Key future directions include uncertainty-aware models that translate detected defects into performance and functionality loss, validated workflows for rapid post-event screening, and integration with network-level decision-making for resilient cities and infrastructure systems.