Trustworthy Human-in-the-Loop AI for Sensing, Diagnosis, and Safety of Specialized Mechanical Systems in Complex Environments

Critical infrastructure and specialized mechanical systems, including transportation and energy infrastructure, offshore platforms, ships, ports, underwater vehicles, and marine machinery, increasingly operate in harsh, uncertain, and safety-critical environments. Their reliable operation depends on more than equipment-mounted sensors: environmental observations, structural and operational data, and human visual, auditory, tactile, and experiential perception all contribute to situational awareness. Artificial intelligence has enhanced monitoring, diagnosis, and autonomous operation, but purely data-driven systems remain vulnerable to distribution shifts, incomplete observations, rare faults, and opaque decisions. At the same time, operator workload, fatigue, attention, and risk perception can affect intervention quality. Integrating machine sensing with human perception, human-state sensing, engineering knowledge, and interactive decision-making is therefore essential for trustworthy mechatronic systems that protect infrastructure, equipment, people, and the environment.This Research Topic aims to advance trustworthy human-in-the-loop intelligence that connects the sensing of equipment, infrastructure, surrounding environments, and human operators with diagnosis, risk assessment, decision support, and safe control. It seeks closed-loop approaches in which multimodal sensor data, physics-based knowledge, human observations, and feedback are jointly used to manage uncertainty and changing operating conditions. Particular emphasis is placed on infrastructure and marine equipment exposed to corrosion, vibration, extreme weather, limited visibility, communication constraints, dynamic loads, and time-critical hazards. Research may address how operator perception, expertise, attention, workload, fatigue, or physiological responses can be measured or incorporated into intelligent systems, and how authority should be adaptively allocated between humans and automation. The overall goal is to enable transparent, auditable, robust, and fail-safe systems that improve diagnostic confidence, support timely human intervention, and safeguard operators, assets, infrastructure, and the environment.To advance trustworthy and human-centered intelligence for complex mechanical systems, we invite Original Research, Review, Methods, Perspective, and Brief Research Report contributions covering, but not limited to:1 Multimodal sensing of infrastructure, marine equipment, and complex environments2 Human perception and human-state sensing using visual, auditory, tactile, physiological, and cognitive cues3 Human-in-the-loop fault diagnosis, prognostics, and health management4 Physics-informed, knowledge-guided, explainable, and uncertainty-aware AI5 Operator feedback, interactive learning, shared autonomy, and adaptive interfaces6 Structural health monitoring and safety assessment of critical infrastructure7 Monitoring, diagnosis, and risk control for ships, offshore platforms, underwater vehicles, and marine machinery8 Digital twins and collaborative human-AI maintenance9 Edge intelligence, resilient communication, and real-time decision support10 Risk assessment, fail-safe intervention, verification, validation, and certification of intelligent mechatronic systems Topic editors: Chuanfei Hu, Hang Shao, Jiapeng Zhang Source: https://www.frontiersin.org/research-topics/85052/trustworthy-human-in-the-loop-ai-for-sensing-diagnosis-and-safety-of-specialized-mechanical-systems-in-complex-environments