HS4U has made significant progress towards automating the assessment of pathogen transmission risk in cruise ships. The team of the University of Thessaly introduced the concept of a closed-loop decision-making model capable of predicting risks associated with the transmission of pathogens on cruise ships and taking actions to mitigate these risks. This is a generic model, applicable to any ship’s compartment. It receives input from sensors monitoring the ship’s areas, analyses the data acquired from the sensors to extract semantic information, and by fusing knowledge from health experts with knowledge extracted from simulation studies, it automatically infers uncertainty-aware decisions that can be communicated to the ship’s crew with natural language to take further actions.

The heart of this artificial intelligence system is a reasoning engine based on fuzzy logic. It encodes data and knowledge using fuzzy sets, which are used to form fuzzy rules. The rules are expressed as “If-Then” statements linguistically quantifying the input and output data, e.g., if the coughing frequency is high in the area and the ventilation rate (air change rate) is low, then the risk for disease spread is high. These rules are combined using mathematical operations to infer decisions with quantifiable confidence. Advantages of this approach include: a) The decisions are tolerant to the uncertainty introduced by the limited and imprecise information available from the sensors; b) the rules are comprehensible by humans; c) the reasoning process is fully transparent and explainable; therefore gaining the users’ trust.

The system’s effectiveness has been systematically investigated for COVID-19 risk assessment using simulations based on validated agent-based models, which can predict the behavior of passengers in actual cruise ship compartments.