Conference Agenda

Port and crane operations are safety critical activities where human error and inadequate training can contribute substantially to operational accidents. Previous research has shown that over 70% of terminal accidents are associated with human factors, such as inadequate training, communication failures and insufficient supervision. A review of 245 studies revealed that only around 2% investigated immersive technologies as a potential approach for reducing accident risks associated with training deficiencies and human factors. Existing safety interventions therefore continue to be focused on primarily on recurrent instruction, procedural compliance, and modifications of training content. Although the adoption of immersive technologies in training has grown considerably in recent years, their use for adaptive workforce training in port operations remains relatively limited.

Building on these findings, this study proposes an adaptive retraining framework for safety critical crane and port operations. Crane-related activities were selected due to their accident relevance, their procedural complexity, and the increasing evidence supporting the use of immersive technologies for training in such environments. The proposed framework focuses on adapting the training process according to trainee performance. Procedural errors trigger different retraining paths, corrective feedback mechanisms, and varying levels of instructional support. Based on the type and frequency of errors, the training workflow can evolve into alternative learning scenarios while also supporting performance based difficulty adjustments. In addition, predefined intervention points allow supervisors to monitor progress and provide guidance when required.

This study focuses on the practical application of mixed reality in training the port operations workforce, moving beyond the development of a retraining framework. Particular attention was given to understanding the opportunities and limitations of using mixed reality for training. This included examining the extent to which operational errors, unsafe actions, and corrective learning situations could be realistically represented in a virtual environment. The developed demonstrator was used to examine how training scenarios can be adapted according to trainees' behaviour and to compare simulated learning experiences with the challenges typically encountered in conventional training settings. Furthermore, the framework generates structured performance and feedback data, which can be used to inform the development of future AI-supported training and feedback systems. Thus, this research contributes to the development of more adaptive, data-driven, learner centred approaches to training the maritime logistics workforce.