Predicting human error before it leads to maritime emergencies

Istanbul, Türkiye - Despite advances in navigation technology, human error still accounts for most maritime accidents globally. The European Maritime Safety Agency (EMSA) reports that human factors played a role in over 80% of marine accidents from 2014 to 2022. Although investigations typically pinpoint the causes of accidents, predicting where errors are likely to occur before a voyage remains a significant challenge.

Researchers at Istanbul Technical University (ITU) developed a predictive modeling framework that identifies potential human error zones during voyage planning. By highlighting these risky areas before departure, it helps shipping firms, maritime authorities, and training groups implement preventative actions, reducing accidents instead of relying solely on post-incident lessons. The research appeared in the journal Safety Science.

Voyage planning is a vital safety task for the ship's bridge team. Before setting sail, officers review navigation charts, check weather forecasts, set up electronic navigation tools, spot hazards, prepare backup routes, and verify all equipment works properly. During the voyage, they keep an eye on traffic, weather changes, and vessel position, making quick decisions to ensure safe navigation. Even small human errors at any stage can lead to collisions, groundings, environmental harm, cargo loss, or injuries.

Previous research has examined specific navigation risks but hasn't systematically evaluated all phases of voyage planning, from initial assessment and planning to execution and continuous monitoring, as an integrated process. This study addresses that gap by offering a comprehensive method to pinpoint where human errors are most likely to occur throughout the entire voyage planning cycle.

The research team created a hybrid analytical framework that merges the Human Error Assessment and Reduction Technique (HEART) with an enhanced Z-number approach. HEART calculates the likelihood of human error in operational tasks, while the improved Z-number method enhances the dependability of expert evaluations by considering both their judgments and the confidence level they have in those judgments. This approach minimizes uncertainty and yields more reliable predictions of operational risk.

The researchers examined 28 specific sub-tasks within the entire container ship voyage planning process, starting from pre-departure preparations, moving through voyage execution, and concluding with monitoring. Nine experienced maritime experts from academia and industry participated in evaluating these operational tasks and identifying the factors most likely to cause human errors. The analysis revealed that the primary risk stemmed from configuring safety parameters in the Electronic Chart Display and Information System (ECDIS). This digital navigation tool helps officers monitor water depths, hazardous zones, and routes. Mistakes in setting safety contours, depth limits, or hazard alerts can lead to false alarms or overlook real dangers, significantly increasing the risk of grounding or collisions.

Other high-risk activities involved interpreting nearby maritime traffic and making maneuvering decisions based on the International Regulations for Preventing Collisions at Sea (COLREGs), verifying the vessel's position through various navigation methods, executing emergency procedures during critical situations, and performing pre-departure inspections of navigation equipment.

Interestingly, many high-risk tasks did not happen during voyage preparation but during the navigation execution and monitoring phases. These activities demand bridge officers to handle multiple information sources at once, such as radar, electronic charts, weather forecasts, vessel traffic, communication systems, and visual cues, while also making quick decisions under operational pressure.

The study also pinpointed several factors that consistently raise the risk of human error. These include inadequate checking procedures, limited hands-on experience, time constraints, stress, fatigue, information overload, and overdependence on automated navigation systems. The researchers suggest that these insights offer practical guidance for shipping companies aiming to enhance bridge resource management and navigational safety.

Rather than replacing experienced officers with automation, the proposed framework supports human decision-making. By forecasting where errors are most likely before operations start, shipping companies can focus training on simulators, update operating procedures, reinforce safety measures, and better allocate resources to the areas with the highest operational risks.

As commercial shipping becomes more digital, with increasingly advanced navigational systems and more complex vessel operations, enhancing maritime safety will rely not only on improved technology but also on a deeper understanding of human performance. The researchers see their framework as a practical step toward proactive safety management, helping organizations identify vulnerabilities early, before accidents happen. Instead of reacting after an incident, predictive human reliability analysis could help prevent emergencies altogether.