Extended Reality to model and mitigate risks of carbon-free ships’ operations using Large Language Models

Research Project, 2026 – 2027

In this project, an innovative framework of risk assessment for those ship operations by integrating hybrid causal logic (HCL) model and oriented objective Bayesian networks (OOBN) will be developed to understand, model and simulate the trigger, evolution and failures/accidents of various risk scenarios, which are defined by AI/machine learning methods using operational experiences from relevant industry fields with huge amount of open reports/database1-5. Due to the incapacity of humans to handle those large database/reports, the functional resonance analysis method analysis (FRAM) from the Large Language Models (LLM) will be developed and implemented to identify various risk factors and corresponding risk evolution process until potential accidents, as well as possible RCOs to mitigate such risks of those accidents. Finally, the above risk scenarios will be developed using extended reality (XR) embedded into Chalmers ship bridge simulators, which help to study and understand the complexity of human factors of seafarers handling risks involved in those carbon-free fuel powered ships. Those human factors will be feed back to the technical development of such alternative fuel power systems and their actual operations. As the forerunner of such innovative development and in-house competences for the operations of carbon-free fuel powered ships, we will expose to more international funding opportunities.