OptiPEx brings passenger-aware services into public transport

OPTIPEX - What it takes to bring passenger-aware services into public transport

The development of automated and intelligent public transport services offers new opportunities to improve passenger experience, accessibility and safety. However, their successful deployment depends on more than technological capabilities. Passenger comfort and trust, data use and protection, or the ability to respond to different user needs are equally important.

This is the focus of the OptiPEx report, “Enablers and barriers for public transport services involving passengers”. The report examines the conditions that can facilitate or hinder the implementation of passenger-aware services, considering technical, ethical and societal, legal, organisational and operational dimensions.

 

Key conclusions

Regulation and compliance – The main legal challenges relate to personal data protection, particularly GDPR restrictions on passenger identification and profiling, as well as uncertainties surrounding the AI Act, certification procedures and approval requirements. Unclear responsibilities regarding data ownership and the sharing of sensitive information with third parties can also delay implementation. To overcome these barriers, the report highlights the importance of integrating legal requirements from the earliest design stages (compliance-by-design) and applying data-minimisation principles. For instance, event-triggered and non-identifiable recordings can replace continuous passenger monitoring, reducing privacy risks and data-management requirements. Furthermore, existing accessibility frameworks such as the European Accessibility Act, provide an established basis for developing compliant and inclusive services.

Ethical and societal aspects – Passenger trust and social acceptance are essential for successful deployment, but may be undermined by concerns about AI transparency, potential algorithmic bias, misuse of personal data and perceived surveillance. Another important barrier concerns the risk of exposing passengers’ individual vulnerabilities, particularly when identifying disabilities or specific assistance needs, potentially resulting in unequal treatment. Technology adoption may also be limited when services require passengers to own smartphones, install applications or register to access certain functionalities. The report therefore emphasises the importance of demonstrating clear and tangible passenger benefits, particularly in terms of safety, accessibility and comfort. Maintaining human control over important decisions is equally important, ensuring that AI supports rather than replaces human judgement. Finally, inclusive design, accessible interaction methods and multilingual support can improve acceptance across different passenger groups.

Safety and security – Safety-related services face significant challenges associated with AI classification errors, including false positives or delayed warnings, which can lead to inappropriate interventions. System failures are particularly critical when services operate in time-sensitive environments requiring immediate communication and intervention from control-centre personnel. To address these risks, the report recommends a human-centred safety architecture, combining continuous human supervision with manual override capabilities, clearly defined escalation procedures and safe-stop mechanisms. Cybersecurity is another essential requirement, with secure communications, encryption protocols and compliance with established cybersecurity standards needing to be integrated from the design stage. In addition, real-time monitoring and early-warning capabilities can improve situational awareness and enable faster responses to potentially dangerous situations.

Technical validation and interoperability – The main technical barriers concern the availability and quality of training and real-world data, the difficulty of validating AI models under realistic operating conditions, and the complexity of integrating solutions into existing transport infrastructures. Variations in lighting, passenger occupancy and environmental conditions can affect system performance, while rare safety-related events remain particularly difficult to detect and validate. Dependencies on external technologies, network connectivity and passenger-owned devices can further complicate implementation. The report identifies scalable and modular architectures, standardised interfaces and the reuse of existing vehicle systems and transport infrastructure as important enablers of interoperability. Additionally, synthetic data generation, data augmentation and open-access datasets can help compensate for limited real-world data. Nevertheless, extensive testing across different operating environments and scenarios remains necessary to ensure reliable performance beyond controlled testing conditions.

Organisational and operational readiness – Even when technically feasible and legally compliant, new services may face organisational barriers that prevent their effective deployment. These include insufficient staff training, limited technical expertise, unclear operational procedures and inadequate preparedness of control centres. Long-term maintenance also represents a challenge, as continuous software updates, diagnostics and performance monitoring can increase operational workloads and lifecycle costs. Furthermore, reliance on external suppliers, third-party integrations and cross-organisational collaboration may create additional dependencies. To facilitate implementation, the report highlights the importance of minimising operational burdens by ensuring compatibility with existing roles and procedures, intuitive functionalities and limited additional requirements. Established partnerships and vendor support can provide the necessary technical expertise and support, while built-in diagnostic tools and remote software updates can simplify maintenance and support long-term service continuity.