Machine learning techniques applied in border crossing

<ns3:p>Background Border control authorities face increasing operational pressure due to growing traveller volumes, evolving security risks, and stricter regulatory requirements. Traditional manual procedures alone are no longer sufficient to ensure both security and smooth passenger flow. The ODYSSEUS project was developed to address these challenges by combining artificial intelligence, computer vision, and multi-sensor technologies into a unified, privacy-respecting solution for border management. Methods The project followed a step-by-step methodology starting with an analysis of user needs, operational workflows, and real scenarios provided by border guards. These inputs were used to define use cases and system requirements. Each ODYSSEUS component was then designed according to the principles of the EU Artificial Intelligence Act, with attention to transparency, robustness, and human oversight. The technologies were implemented, tested in controlled environments, and gradually integrated into a modular platform. Finally, the complete system was validated in three pilot locations representing land, sea, and rail border environments. Results The ODYSSEUS platform successfully combined several advanced tools, including digital travel credential verification, behavioural authentication, faceless people counting, unmanned aerial vehicle-based vehicle inspection, and artificial intelligence-supported X-ray screening. All components were integrated into a single decision support interface that provided real-time information and alerts. The pilots demonstrated that the system could support officers by improving situational awareness, reducing manual workload, and enabling faster processing, while maintaining strict compliance with data protection requirements. Conclusions The project shows that a modular and artificial intelligence-enabled approach can enhance border control operations without compromising privacy or regulatory obligations. The successful demonstrations in three different environments confirm that the ODYSSEUS technologies are mature, functional, and suitable for further deployment and scaling. The work provides a practical foundation for future integration of artificial intelligence systems in European border management.</ns3:p>