Information Requirements and Legal Framework for Multimodal Transport System Coordination

Abstract

Background: In multimodal transport the interplay of coordination methods and legal requirements is a challenging task. To address the latter, a combined approach for the coordination of a multimodal passenger transport system in accordance with European data protection law is required. Method: As a first step the paper analyses coordination related delays and outlines a combined optimisation problem. The problem formulation spans the strategic, tactical and operational level, to identify information requirements depending on coordinationmechanisms. The European legal systemregularly sets a pioneering standard, often serving as a model for other countries. Additionally, European data regulations frequently influence international data flows, as references to European traffic standards are often indispensable. To ensure compliance with data protection legislation, in a second step, this paper analyses the European Union’s legal framework for the protection of personal and non-personal data. Result: A respective system architecture for the integration of selected methods is proposed and the resulting analysis outlines the legal requirements for data usage under the Data Governance Act (DGA) and the General Data Protection Regulation (GDPR). Conclusions: Achieving a sustainable and efficient transportation system requires a balanced integration of advanced data-driven solutions and legal strategies, ensuring system efficiency and compliance with EU protection laws.

1. Introduction

The European Union (EU) is in the process of transforming its transportation system, as outlined in the European Green Deal (EGD). The transport sector is a significant contributor to the EU’s greenhouse gas emissions, accounting for 21.3 percent of overall emissions in 2021 [1]. With the EGD, the EU has set the ambitious goal of reducing greenhouse gas emissions to net-zero by 2050. To achieve this target, the EU aims to reduce emissions in the transport sector by 90 percent. Consequently, the transport system must be enhanced not only in terms of efficiency and cost-effectiveness but also to ensure compliance with strict environmental and legal standards while remaining socially fair and equitable [2]. To support this transition, the European Commission presented the Sustainable and Smart Mobility Strategy, accompanied by an action plan. This strategy is structured around three main pillars: making transportation more sustainable, increasing the availability of multimodal transport options, and establishing incentives to drive the transformation [3].

In alignment with this strategy, the Sustainable and Energy-Efficient Aviation (SE²A) research cluster develops new technologies for a sustainable and eco-friendly air transport system (ATS) [4]. As part of an integrated system, these technologies will have impacts on the flow and operating processes regarding the hinterland transportation of travellers, who access and egress airports during their journeys within the EU. These journeys may consist of different modes of transportation (multimodality), such as car, bus, rail micromobility, walking and include at least one air transport segment and change of modes. To render such a system to be efficient and robust and to reach the EU’s vision of a 4-h-door-to door goal, the sharing of information between different stakeholders, like transport operators, infrastructure providers and passengers becomes essential [5,6].

From a methodological perspective, this raises the question of what data is actually needed and how it can be accessed. The EU aims to enhance the availability of data and promote a robust data ecosystem through the adoption of the Data Governance Act (DGA). Moreover, data protection requirements present significant challenges. It is therefore necessary to analyse the interaction between the DGA and the General Data Protection Regulation (GDPR). EU law plays a central and influential role in data protection, setting pioneering standards [7]. For instance, the GDPR has directly inspired the development of Brazilian data protection laws. However, other nations are aligning their data protection standards with those mandated by the GDPR in Europe, particularly for data concerning European citizens (e.g., India, California) [8,9]. Additionally, it is crucial to note that when the data of subjects in the EU is processed, European data protection law applies, as outlined in Article 3 (2) GDPR [10]. Non-compliance with the requirements of the GDPR can lead to severe penalties, including fines of up to EUR 20,000,000 under Article 83 GDPR, or, for companies, up to 4% of their total worldwide annual turnover from the preceding financial year [11]. This extraterritorial reach further underscores the EU’s role as a global model and benchmark in data protection. Consequently, it becomes evident that an analysis of European data protection law is of significant relevance not only within Europe but also on a global scale.

The structure of this publication integrates both technical and legal perspectives, reflecting the growing importance of the link between law and technology, particularly in light of ongoing digitalisation. Both sections serve as standalone parts, that respectively provide technical aspects for juristic readers and juristic considerations for technical readers. The technical section details the state, sources of delay and challenges within the transport system. Followed by the definition of a problem specific transportation graph, a review on information requirements and a proposed structure of suitable methodologies to optimise the transport system.

Parallel to this, the legal aspect of the paper focuses on the EU’s data strategy. It evaluates the handling of protected data, detailing aspects of personal and non-personal data protection and the aspects of data anonymisation.

Last, the paper draws these strands together in a comprehensive conclusion. This joint conclusion synthesises the insights from both the technical and legal perspectives, providing a holistic view of the future of transportation within the EU.

2. Technical Aspect

This section analyses available transportation data from current initiatives and provides an overview of the passenger transportation structure within the EU. We particularly focus on primary sources of delays and classify transportation coordination challenges. Methods regarding the latter are linked to information requirements to obtain an overarching structure.

2.1. Transportation Related Initiatives by the European Union

The primary EU initiatives that consider data as an vital aspect are Flightpath 2050 [5], DATASET2050 [12], IMHOTEP [13] and Mobility4EU [14].

Flightpath 2050, envisioned by the European Commission, sets the long term targets for the European aviation sector by 2050. The focus is set on an efficient and seamless transport while reducing the environmental impact. A key objective is that “90% of travellers within Europe are able to complete their journey, door to door, within four hours”. Additionally, it aims for a 75% reduction in CO₂ emissions per passenger kilometre, a 65% decrease in perceived noise emissions, and a 90% reduction in NO_x emissions from aircraft [5].

Supporting these goals, the DATASET2050 project employs a data-driven, passenger centric approach to analyse how supply of transport options may match the current and future demand, and to identify weak areas and bottlenecks in the transport chain. Based on those insights, transport concepts and possible solutions for current and future limitations toward achieving the four hour door-to-door goal are defined [12].

The IMHOTEP project aims for an integrated multimodal transport system with airports as connecting nodes. It takes into account the hinterland operations for real time collaborative decision making at the airport. To achieve this, methods to analyse big-data retrieved from the door-to-door transport are used. Based on the latter, forecasts of airport passenger flows and the assessment of the effects on the airport and hinterland operations are extracted. The methods and tools are validated on two case studies at the London City and Palma de Mallorca airport [13].

Lastly, the Mobility4EU project provides a vision with requirements, challenges and opportunities. It’s focus lies on a user centred transportation system within the EU in 2030, paired with an action plan for the implementation of that vision [14].

2.2. Overview of Passenger Travel in the EU Transportation Network

In this section, we explore passenger travels within the transport network of the EU, supported by the results and information provided by the previous EU initiatives, in order to establish a basic understanding of transport segments, modes and their sources of delays for further analysis.

This transport process can be divided into several distinct segments, starting and ending at a door, referred to as door-to-door (D2D). Understanding these is essential for the overall network coordination, where at least two local decision making problems are synced, and for addressing specific issues that may arise in each segment. The segments are as follows [12]:

• Door-to-Kerb (D2K) represents the section from the starting point of the journey “Door” to the point where a passenger enters the airport at the “Kerb”. This transportation segment may include different modes of public and private transportation, such as car, micromobility, bus or train. In addition to this multimodality, the transport modes may be changed during this segment, resulting in intermodality.
• Kerb-to-Gate (K2G) represents all processes required for the departure within the airport, like check-in, boarding, walking, luggage handling and security checks. This segment is also influenced by the buffer each passenger plans, to be certain to arrive in time for the departure.
• Gate-to-Gate (G2G) represents the segment from the start of boarding to the beginning of deboarding off the plane, including processes like flight and taxiing.
• Gate-to-Kerb (G2K) represents all processes required upon arrival at an airport. Similar to the K2G, but distinct in the way that there exist no planned buffer from the traveller but different processes like immigration checks become relevant.
• Kerb-to-Door (K2D) represents the last segment in the transportation chain. It begins when the passenger leaves the airport over the “Kerb” and moves to his destination, the “Door”. Similar to the D2K segment, it may allow for different modes of transport.

Based on the results reported in Dataset 2050 [12], the mean time per segment, is 33 min for D2K, 114 min for K2G, 150 min for G2G, 31 min for G2K, and 28 min for K2D, with 5 min as estimated acceptable error due to the stochastic simulation and dataset. As a result, the mean D2D time is 356 min. In total, only around 10% of passengers are able to complete the D2D in less than 4 h or 240 min and 90% take less than 460 min.

2.3. Coordination Related Sources of Delay in the Transportation Network

Within each segment, there exist sources of delay, which increase the time spent per segment. Here, we focus on specific sources related to coordination.

For D2K/K2D, Paulsen et al. [15] showed that uncertainty in public transport services results in variation of travel time. The authors of [12] reported that the time of day has an impact on travel speeds and therefore an influence on the time spent on travelling. In [16], it was observed that seemingly short dwell times on station platforms can sum up to large delays. In [17], the authors found that the primary source for delays in a public transport system appears within the stations. In [18], boarding and alighting was shown to affect dwell time, especially for bus transport. Higgins et al. [19] and Haywood et al. [20] showed that congestion’s in private and public transport have a negative influence. Similarly, bottlenecks within the transport network also contribute to this negative impact [21]. With a focus on coordination, in [22] the authors reported that missed connections due to a lack of coordinated delay management at intermodal transport hubs can lead to increased travel time. Similarly, in [23] the reliability between modes in an intermodal transport system showed impacts on travel time. The authors of [24] stated that delays in communication regarding flight arrival and departures have a negative effect. In [25], an improved timetable coordination for public transportation modes allowed for a reduction in waiting time and delays. Moreover, Gallotti and Barthelemy [26] stated that an average of 23% in travel time in multimodal transports within the UK was lost in connections between modes from a lack of synchronisation. Last, the social impact was discussed in [27].

For K2G/G2K, Hsu et al. [28] analysed the effects of delay propagation onto the following processes within airport passenger terminals.

Regarding G2G, the authors of [29] showed that the implementation of a coordinated strategy regarding ground transport and alternative flights can lead to a 70% reduction in delays. Castaing et al. [30] reported that gate blockage from a variation in flight times and departures results in delays. In [31], delay caused by ground holding of flights at the departing airport was found to induce additional delays further down the transport chain.

Combined, there exist various causes in access and egress passenger behaviour and transport mode availability that induce the inequality in travel time. The most prominent one is the inclusion of additional buffer time due to the fear of missing a flight within the K2G segment [12].

In order to gain an understanding of how to approach coordination related delays within the transportation network, we summarise them in Table 1, highlight their influence on coordination and provide a reasonable approach to the coordination problem.

Table 1. Analysis of delaysand respective impacts in multimodal transport systems and possible remedies.

Delay	Source	Influence on Coordination	Approach
Uncertainty in public transport services	[15]	Causes variations in travel time, affecting the efficiency in synchronisation between different modes of transport and causing additional buffer time in the route planning.	Determine the probability improvement of the estimated travel time adherence.
Time of day impact on travel speeds	[12]	Predictably affects travel time and may effect schedule adherence and scheduling through variation in traffic density.	Delay mitigation via dynamic scheduling and predictive route adjustments.
Dwell times at stations	[16]	Leads to unpredictable dwell times due to passenger behaviour variability at stations, causing cumulative delays.	Simulate the impact of platform management and information sharing on reducing dwell time variance to enhance schedule reliability.
Delays in stations	[17]	Disrupts overall transportation coordination as a result of unforeseen delays in stations, leading to additional delays due to missed connections.	Integrate automated passenger data collection into simulations to enhance accuracy in modelling station delays and improve operational coordination.
Boarding and alighting times	[18]	Negatively impacts average operating speed and schedule adherence, causing bus delays due to frequent stops and passenger related factors.	Simulate passenger dwell time delays and assess their effects on the transportation network.
Congestion’s in public and private transport	[19,20]	Causes widespread delays, impacting the transport network.	Alleviate congestion delays by determining optimal route, schedule and mode through real-time adjustments, via real-time data sharing.
Bottlenecks within the transport network	[21]	Creates critical sections with traffic flow limitations, causing repeated and in cases predictable delays.	Identify bottlenecks and evaluate methods to avoid reaching the limit at which the flow starts decreasing.
Missed connections due to a lack of coordination at intermodal stations	[22]	Leads to increased waiting times and inefficient changes between modes.	Analyse intermodal hubs and identify waiting times between modes, passenger information requirements and systematic problems that result in delays.
Reliability issues in intermodal transport systems	[23]	Affects overall travel time and passenger satisfaction, as low reliability also leads to an earlier start of the passenger journey.	Determine how changes in reliability of certain modes influence travel time and service reliability.
Delays in communication regarding flight arrival and departure changes	[24]	Affects primarily downstream transport coordination leading to missed adjacent transportation.	Determine the effect of communicating flight data to the hinterland transportation system.
Lack of synchronisation and timetable coordination between modes	[25,26]	Incorporates long waiting times due to changes between modes, prolonging the initial travel time, as no information is exchanged between the stakeholders.	Evaluate the benefit of scheduling across transportation modes and evaluate if dynamic scheduling across modes is viable.
Gate blockage at airports	[30]	Causes delays in boarding and deboarding due to variations in arrival and departure times of airplanes.	Analyse integration of gate on flight control.
No coordination for alternatives for delayed flights	[29]	Not incorporated alternatives to flights with long delays into the passenger journey.	Incorporate alternative transport modes to a chosen flight that has been delayed.

2.4. Information Requirements of Approaches for Transportation Coordination

The outlined delays lead to complex challenges that span logistical, technological, and legal domains. In the legal domain, requirements are imposed on topology design for mobility providers, which in turn protect individuals by ensuring movement within legally compliant spaces. The complexity of integrating different transportation modes with distinct operational characteristic was described in works like [32], where the aspects of intermodal and multimodal transportation and the need for coordination between the transportation methods was noted.

The literature points towards data and information as a critical component for the technological transitions required for sustainable transportation systems. In particular, researchers highlighted the need for solutions that integrate various transportation modes, while adhering to legal standards and leveraging modern technologies [33,34]. The role of Big-Data was particularly emphasised in this context. The authors of [35] discussed how Big-Data can be initialised in enhancing logistical operations through real-time data analytics and predictive modelling, thereby optimising routes and improving coordination. In [36], inefficiencies in multimodal transport were found to be due to a lack of coordination on the operational level, compromising the possible advantages. To be able to treat the overall problem within the system, we have to differentiate in between a passenger centric and a system centric perspective, resulting in distinct formulations of the optimisation problem. From the system centric perspective, the objective is to minimise overall costs by incorporating key performance indicators (KPIs) related to system performance. The costs are determined by a set of factors, with weights applied to each based on its importance. These factors may include operational costs, delays, or resource utilisation, and the total sum of weights must be equal to one, ensuring all factors contribute proportionally to the objective.

From the passenger centric perspective, the goal is to find the cost minimal path, which reflects individual preferences such as comfort, convenience, or the cost of transportation. Each passengers optimisation problem considers KPIs that may vary from those of the system and other passengers, in order to reflect personal preferences.

The combined optimisation problem then seeks to balance the systems needs with the individual preferences of passengers.

A balancing factor is used to adjust the weight between the system centric and passenger centric approaches, ensuring that both perspectives are considered in determining the optimal solution. This balance allows for an adaptable approach that can prioritise the system, passengers, or a compromise between both, depending on the requirement.

To address these challenges, methods such as Distributed Model Predictive Control (DMPC) and Reinforcement Learning (RL) are of particular interest in transport coordination [37,38,39,40,41,42], which in turn require information for coordination. Transportation systems in particular exhibit a distributed nature with several stakeholders, including transport operators, infrastructure providers, and passengers, each acting based on local criteria and constraints. Hence, the exchange of information between stakeholders may be necessary to solve the combined optimisation problem. We focus on the type of information available to support such coordination, which we then match with possible coordination methods in the following section.

In [43,44], smart cards provide personal data such as name, age, nationality, and home/work addresses, along with travel specific data, like luggage details, destination and public transport card ownership. Additional data includes information on boarding time and location. The authors of [45] summarised that retrieval of position of vehicles can be done using the Global Positioning System as part of an automated vehicle location, whereas passenger counts may be estimated using automated passenger counting systems. With the use of video footage, Guo et al. [46] showed that waiting times, arrival and departure times can be determined. In [47], the authors deducted the optimal frequency on each route with constraints regarding fleet size, minimum required frequency and capacity of route for transport operators.

The authors of [48] specified operational cost and passenger revenues as important indicators for transport operators. Additionally, the service reliability for the infrastructure provider regarding punctuality and regularity of transport was shown to be an important factor. Similarly, reliability of service plays a significant role for passengers as it influences travel time. In [49], it was summarised that passengers typically value reliability, speed and comfort of a transport system. A framework for calculating service reliability was provided in [50], whereas speed is typically calculated from waiting and in-vehicle time and comfort is determined from the average standing time during travel [48]. In [51], the authors outlined the six criteria transit time, punctuality, time and spatial offer, comfort, transportation cost and impact of mass transit system on city environment to assess the quality of a mass transit system from a passenger viewpoint. They are further ordered into sub-criteria, detailing the criteria and separating the assessment into a time and a point scale.

Based on an expert survey, the authors of [52] extracted weighted criteria with sub-criteria for passengers in a transport system in Tehran. The main criteria with weights are comfort (0.223), accessibility (0.205), time (0.156), payment (0.165), safety and security (0.184) and environmental impact (0.067).

As part of the DATASET2050 [12] project, twelve passengers metrics reflecting possible future passenger needs have been derived, these include options and control, reliability, personal data security and protection, meaningful use of travel time, journey time, safety, family friendliness, environmental friendliness, age appropriateness and accessibility, digital usage, affordability and available network of connections.

In [53], a case study at Palma De Mallorca Airport and London City Airport retrieved multiple datasets. These included airport layout and processes (such as luggage information, check in), flight schedules and attributes (number of passengers and destination), passenger security control information (flight information and location in airport), passenger surveys (airport access and egress behaviour and behaviour inside the airport), airport flow data (passenger flow that accesses and egresses from the airport via public transport), mobile phone data (position and sociodemographic information, behaviour during disruptions), camera based counting (number of vehicles and passenger accessing and leaving the airport by a certain transport mode), and WiFi data (amount of users at certain areas within airport).

Factors describing objectives for airports like low operational cost, luggage processing, quality of service and access to multiple modes of transport were described in [54,55].

For a more general overview of the available data within Europe, the Mobility Data Space provides information regarding traffic, live traffic flow, infrastructure, public transport, micromobility sharing and more [56]. Specifically, it includes information regarding the location and availability of parking spaces for each street bound mode, road information (speed limit, tolls), public transportation delays and vehicle data, origin-destination-matrices and multi-modal accessibility.

Table 2 summarises the above with regards to states, constraints and objective/cost functions. Note that in order to be fit for use in coordination methods, the listed types reflect a filtered version, which needs to be gathered by merging underlying data into broader factors.

Table 2. State, constraints, and objective/cost functions for transport operators, infrastructure providers, and passengers.

	Transport Operator	Infrastructure Provider	Passenger
Stakeholder	Rail, bus, taxi, ridesharing, car rental, micromobility companies, airlines	Airports, rail stations, bus terminals, micromobility hubs, government agencies	Individual
State	Position, passenger count, operational status	Passenger count and flow, person activity (waiting, walking, queuing), arrival and departure times	ID, age, nationality, home/work addresses, luggage, destination, transport card ownership, position, mode of transport
Constraints	Fleet size, minimum required frequency and capacity, vehicle limitations	Layout, available transport modes for arrival and departure, person and vehicle capacity	Financial spending capability, group or single traveller, environmentally conscious decision making, accessibility restrictions
Objective/Cost Function	Optimal frequency, operational cost and passenger revenue, service reliability	Service quality, schedule, operational cost and passenger revenue	Service reliability, safety, comfort, travel time, cost, accessibility, environmental impact

2.5. Methodology

To achieve optimal control, coordination and communication between the stakeholders of a the transport system is required. As these stakeholders act on different time scales, we divide the system into the strategic, tactical and operational layer, where each layer is utilised for a different planning horizon and possesses different characteristics. Coordination and communication then exhibits both a horizontal (within layers) and vertical (between layers) structure [47,57]. A representation of that structure is shown in Figure 1.

Figure 1. Strategic, tactical, and operational layer of the transport system, with the feedback structure between them.

The selection of specific optimisation methods for each layer depends on the use case and its inherent properties. Therefore, an overview of specific methods is not feasible. Instead, we provide a generic approach to the problem by first describing the properties of each layer and then offering a high-level classification of optimisation methods. Suitable methods may be categorised into machine learning, deterministic, heuristic and stochastic optimisation. Machine learning methods aim to improve performance measures through training and are further divided into supervised learning, unsupervised learning, and reinforcement learning [58]. Heuristics are employed when no polynomial-time algorithm can solve an optimisation problem, searching the feasible region for a solution that may not be optimal, but is computationally efficient for complex or NP-hard problems [59]. Stochastic methods incorporate uncertainties into the model and optimise the expected performance, while deterministic methods rely on fully known models, providing replicable and unequivocal solutions [60,61]. To prime any of these methods, hyperparameter optimisation can be applied, which focuses on finding the optimal hyperparameter configuration [62,63]. Simulation can be used to address dynamics within transport problems, and to integrate the various layers with each other [64].

Incorporating the overall approach from Figure 1 into one overreaching problem, we obtain the structure illustrated in Figure 2.

Figure 2. Hierarchical system architecture with integrated strategic, tactical, and operational layers, including information flow and optimisation methods.

Moving through the layers, the strategic layer is typically modelled as a graph problem, followed by the tactical layer, implemented as a macroscopic model, depending on aggregate variables, and finally the operational layer, which focuses on microscopic models that include individual vehicle dynamics. Alternatively, a mesoscopic model could be used for both the tactical and operational layer, balancing computational efficiency with behavioural detail through probability distribution functions that represent the behaviour of the individuals [65].

In more detail, the strategic layer is used to solve the design problem of transport network graph by providing the network structure and weights for the tactical layer in form of parameters and the availability of schedules and services. This layer is responsible for long-term planning and involves considerations of infrastructure development, capacity planning, and regulatory compliance. It takes the network structure, constraints, objective functions and KPI as an external input, enhanced by feedback from the tactical layer.

On the tactical layer, the focus shifts to mid-term planning and optimising the use of the network structure provided by the strategic layer. This layer represents the transport systems dynamics in macroscopic detail and handles mid-term decisions such as route planning and resource allocation. The tactical layer translates the strategic goals into actionable plans, providing an operating point for the operational layer in the form of a cost-minimal path for each passenger and for the system.

The operational layer handles real-time decision making at a microscopic level, addressing immediate responses to disruptions such as congestion, demand fluctuations, or technical issues within the transport network. This layer is highly dynamic and relies on continuous input from the tactical layer to adjust plans as needed. Operational activities that necessitate re-planning are fed back into the tactical layer, enabling adjustments to the planning processes. The results of these tactical adjustments are further used to validate and refine the strategic design.

The coordination between these layers is crucial for the systems overall effectiveness. There is both horizontal coordination within each layer and vertical communication between layers. Feedback loops are established where data and outcomes from the operational level inform tactical adjustments, which in turn lead to strategic realignments. This bidirectional flow of information ensures that the system remains adaptive, responsive, and capable of meeting its objectives in a complex and dynamic environment.

2.6. Summary

The technical section provides an overview of EU transportation, focusing on current EU projects like Flightpath 2050 and DATASET2050, which consider data as a critical factor in transportation and set goals for the EU transportation sector. One such goal is the 4 h door-to-door travel time, which is currently achieved by only 10% of travellers, with the mean journey time being around six hours. The passenger journey is further divided into five segments. Based on this, the information requirements for coordination methods and the current challenges in the transportation system are discussed. This is then summarised for the transport operators, infrastructure providers and passengers, who possess certain states, constraints and objective/cost functions, depending on the respective stakeholder.

The influence of certain delays on the coordination of the system is analysed and suitable approaches defined. We further propose a combined optimisation method that balances system centric objectives, such as operational cost minimisation, with passenger centric objectives, like comfort and convenience.

To achieve coordination, a multi layered architecture, with strategic, tactical, and operational layers optimising long-term planning and design, mid-term route planning and task allocation, and short-term control actions towards disturbances is introduced. Feedback between the layers ensures adaptability, enabling improved coordination and overall system performance. Machine learning, heuristic, stochastic, and deterministic methods are considered suitable for the methodology.

3. Legal Aspect

The European Commission aims to implement targeted political measures and strategic investments through its data strategy by 2025 [66]. The goal is to cultivate a data economy that harmonises with European values, fundamental rights, and regulations [66]. Key elements of this strategy include improving access to data and promoting responsible handling of data [66]. The objective is to create a unified European internal market for data, ensuring that companies have real-time access to essential data while adhering to the legal framework [66]. The administration, maintenance, and reuse of data are facilitated to ensure trust in data-sharing procedures through an internal data governance and data policy [67,68].

Among other measures the transition to a strengthened union-wide digital architecture, as part of ensuring a seamless digital landscape, will be ensured by the increased use of open-source software, the utilisation of cloud technology, the realisation of cross-agency pilot projects for modernisation and innovation, and additional initiatives [66]. Furthermore, concepts for the future standard interoperability of IT systems and corresponding deployment systems are developed, and the involvement of end-users in the design of future systems is promoted for a positive user experience [66].

In order to apply the methodology to make valid and optimal decision predictions, a large amount of data is required, as shown in Table 2. Data such as the age or nationality of a passenger or data concerning travel preferences, the location of aircraft, buses or trains, but also arrival and departure times, the emissions of the means of transport are all subject of data regulation. This raises the question of the extent to which this data can be accessed and whether it may be processed. With regard to access to data, the DGA (Regulation (EU) 2022/868 of the European Parliament and of the Council of 30 May 2022 on European data governance and amending Regulation (EU) 2018/1724 (Data Governance Act), OJ L 152, 3.6.2022, p. 1–44.), which came into force on 23 June 2022, is of particular relevance. Apart from this, other regulations such as the GDPR (Regulation (EU) 2016/679 of the European Parliament and of the Council of 27 April 2016 on the protection of natural persons with regard to the processing of personal data and on the free movement of such data, and repealing Directive 95/46/EC (General Data Protection Regulation), OJ L 119, 4.5.2016, p. 1–88.) and the national data protection laws of each member state must be taken into account. Other regulations, such as the Data Act (Regulation (EU) 2023/2854 of the European Parliament and of the Council of 13 December 2023 on harmonised rules on fair access to and use of data and amending Regulation (EU) 2017/2394 and Directive (EU) 2020/1828 (Data Act), OJ L, 2023/2854), the Trade Secrets Directive (Directive (EU) 2016/943 of the European Parliament and of the Council of 8 June 2016 on the protection of undisclosed know-how and business information (trade secrets) against their unlawful acquisition, use and disclosure, OJ L 157, 15.6.2016, p. 1–18.), the implementing regulation of the commission (EU) 2023/1385 (Commission Implementing Regulation (EU) 2023/138 of 21 December 2022 laying down a list of specific high-value datasets and the arrangements for their publication and re-use, OJ L 19, 20.1.2023, p. 43–75.) which laying down a list of specific high-value datasets and the arrangements for their publication and re-use, also define further requirements for the use of data.

In this context, the question arises as to what extent data protection regulations make access to certain data more difficult. In this respect, the connection between the DGA and the GDPR must be evaluated. Therefore it is important to provide an overview of the legal framework for data use requirements. The analysis is constrained to the European legal framework, as an exhaustive examination of the entire corpus of legal provisions would exceed the scope of this essay.

3.1. The Importance of Data Protection Law for Technology

Nowadays, technical innovation is hardly conceivable without the use of data. In this respect, the premises of ‘privacy by design’ and ‘privacy by default’ were already established in the 1990s [69]. According to this premise, the risk potential of data processing systems and processes should be proactively prevented through technical design [70]. This realisation is also reflected in Art. 25 GDPR. Taking into account the state of the art, the cost of implementation and the nature, scope, context and purposes of processing as well as the risks of varying likelihood and severity for rights and freedoms of natural persons posed by the processing, the controller shall, both at the time of the determination of the means for processing and at the time of the processing itself, implement appropriate technical and organisational measures, such as pseudonymisation, which are designed to implement data-protection principles, such as data minimisation, in an effective manner and to integrate the necessary safeguards into the processing in order to meet the requirements of this Regulation and protect the rights of data subjects.The controller shall implement appropriate technical and organisational measures for ensuring that, by default, only personal data which are necessary for each specific purpose of the processing are processed. That obligation applies to the amount of personal data collected, the extent of their processing, the period of their storage and their accessibility. In particular, such measures shall ensure that by default personal data are not made accessible without the individual’s intervention to an indefinite number of natural persons. Effective data protection can therefore only succeed if it is taken into account from the outset in the development of technical systems [71].

Although the provision of Art. 25 GDPR is a purely procedural provision, it is regularly in the interest of developers to take data protection requirements into account from the outset [70]. Subsequent changes to adapt to data protection requirements are usually associated with a high level of effort and therefore also with financial losses [72]. There is also the consequence of Art. 83 (4) and (5) GDPR. According to this, anyone who violates the provisions of Art. 25 GDPR must pay a fine of up to EUR 10,000,000 or, in the case of a company, up to 2% of its total worldwide annual turnover in the preceding financial year. On the other hand, if the provisions of Articles 5, 6, 7 and 9 GDPR have been breached, the fines can reach up to EUR 20,000,000 or, in the case of a company, up to 4% of its total worldwide annual turnover in the previous financial year. It is therefore important to take data protection issues into account in the technical design from the outset.

3.2. The Use of Legally Protected Data

In the EU, security and data protection are closely linked to the implementation of the GDPR. A fundamental determinant of this regulatory framework, as set out in Article 2 (1) GDPR, is its exclusive application to personal data. Consequently, a meticulous examination of data becomes imperative to discern its classification as either personal or non-personal.

Article 4 (1) GDPR provides a comprehensive definition of personal data, characterising it as any information pertaining to an identified or identifiable natural person, herein referred to as the "data subject". The crux of the matter centres on the fundamental question around the core query of whether an individual can be reasonably identified. According to the stipulations of Article 4 (1) GDPR, an identifiable natural person is one who can be identified, directly or indirectly, in particular by reference to an identifier such as a name, an identification number, location data, an online identifier or to one or more factors specific to the physical, physiological, genetic, mental, economic, cultural or social identity of that natural person. The nuanced language employed in the regulation underscores its intent to encompass a diverse array of scenarios. It embraces all forms of information—both objective and subjective—interconnected with a specific individual [73]. Therefore, if information allows for a direct inference or statement about an individual, it attains the classification of personally identifiable information [74,75].

In the realm of established legal doctrine, the presumption of identifiability is upheld when the combination of information can lead to a personal reference [76,77,78,79]. Identifiability is duly acknowledged when an individual can be distinctly discerned from others in a group without necessitating further differentiation [80]. Crucially, the personal nature of data endures as long as the data subject remains identifiable by the data controller [73]. This underscores the intrinsic dependency of data’s personal nature on additional knowledge accessible to the data controller [81]. According to Article 4 (7) GDPR, the term ‘controller’ refers to the natural or legal person, public authority, agency, or other body that, either independently or in collaboration with others, establishes the purposes and methods of processing personal data. For those who do not possess this additional knowledge, data is anonymised and therefore does not qualify as personally identifiable information [81]. Recital 26 of the GDPR elucidates that in order to ascertain whether an individual is identifiable, consideration should be given to all the potential means that could reasonably be employed to identify them, whether by the controller or by another party [76]. If the identification process would require a disproportionate amount of time, effort, or resources, the individual should not be considered identifiable [76]. This implies that anonymisation must be sufficiently robust to render identification practically impossible for any party that could reasonably access the data [76].

Data swiftly acquires a personal dimension, vividly illustrated by instances like location information. Inherently devoid of personal particulars, locations fundamentally lack an inherent personal connection, merely serving as geographical indicators for objects [76]. However, the nuance emerges when considering the potential surveillance of a taxi driver, contingent upon knowledge of the driver’s identity [76].

This perspective extends seamlessly to other categories of data. If the data custodian has a comprehensive view of transport schedules, departure and arrival times, coupled with an awareness of the identities of the respective drivers, then these datasets transition seamlessly into personally identifiable information about the individuals involved. In essence, this highlights the complex interplay between data and personal context, where the potential for personalisation comes not from the data itself, but from the additional layers of contextual knowledge held by those who manage the information. Consequently, the scope of the GDPR is quickly opened up, as a personal reference can be established.

3.2.1. The Protection of Personal Data

In the event that the GDPR is applicable, the handling of personal data is governed by the stipulations outlined in Articles 5 and 6 GDPR. Consequently, adherence to the principles set forth in Article 5 GDPR is imperative, necessitating that the processing be conducted lawfully, grounded in good faith, transparently executed, driven by a specific purpose, and characterised by factual accuracy. The same applies to optimisation processes in traffic management systems.

The term “processing”, as elucidated in Article 4 (2) GDPR, encompasses a broad spectrum of operations, whether automated or not, that pertain to personal data. Article 4 (2) GDPR provides illuminating examples of these operations, including but not limited to collection, recording, organisation, structuring, storage, adaptation or alteration, retrieval, consultation, use, disclosure by transmission, dissemination, or any other form of provision, alignment or combination, restriction, erasure or destruction. Consequently, especially in the realm of utilising personal data within the domain of Artificial Intelligence, the concept of processing, as delineated in Article 4 (2) GDPR, becomes evident, as data is, at a minimum, read and utilised.

The core tenet of data protection is the lawful processing of information. This commitment is explicitly outlined in Articles 6 and 9 GDPR. This is a fundamental principle, asserting that any data processing must be based on a clear legal basis [82]. While this regulatory technique may be atypical in civil law, the caveat applies equally in private law [83]. Article 6 (1) GDPR allows data processing only if at least one of the specified legal bases is applicable [82]. The legal bases listed in this article are exhaustive and constitute the sole basis for lawful data processing [82].

As a fundamental scenario Article 6 (1) GDPR establishes, when the data subject has given consent, the lawfulness [84]. In principle, when the data subject has given consent, comprehensive data processing is permissible. The conditions for effective consent arise from a comprehensive examination of the provisions of Article 4 Number 11, Article 6 (1) (a), Article 7 and 8 GDPR, along with the general principles of data protection law [85]. However, a common challenge with consent typically lies in the specification of the purpose, as it must be explicitly formulated [84]. Furthermore, the privacy statement must provide clear references about the nature and scope of data processing [84]. Nevertheless, the precise objective of data processing in a traffic management system cannot be identified in advance for a specific potential consent scenario.

Without consent, the processing of personal data is only lawful when it is necessary for the performance of a contract or pre-contractual measures, for compliance with a legal obligation, to protect vital interests, for the performance of tasks carried out in the public interest, or for the protection of legitimate interests [86]. In this regard, it is fundamentally conceivable that the processing of personal data, even without consent, could take place within the framework of traffic management systems under Article 6 (1) (b–f) GDPR [85]. However, this is constrained by Article 9 GDPR. According to this article, the processing of sensitive personal data, including racial or ethnic origin, political opinions, religious or philosophical beliefs, trade union membership, genetic data, biometric data for the purpose of uniquely identifying a natural person, data concerning health or data concerning a natural person’s sex life or sexual orientation, must always be preceded by the explicit consent of the data subject, or there must be a lawful basis for the processing [87]. Furthermore, additional requirements for processing are stipulated under Article 9 GDPR [88]. Biometric data is important to emphasise in this context. According to the legal definition in Article 4 No. 14 GDPR, this includes, for example, fingerprints, vein or iris identification [88]. The aforementioned data is already being utilised in several airports, including Dubai, to streamline security checks [89].

Additionally, there is the purpose limitation set out in Article 5 (1) (b) GDPR. This provision limits the legal basis to the specific purpose to be determined. [82]. The combination of these legal principles ensures that data processing is not only carried out on a clear basis, but is also strictly limited to the predefined purposes [82]. The specified purpose must be unequivocal [83,90,91]. Therefore, the purpose should not be too broadly defined or ambiguously described [83,90,91]. The admissible processing of data is determined by the pursued purpose [91,92].

3.2.2. Anonymisation and Pseudonymisation of Data

Anonymisation and pseudonymisation of data are therefore of crucial importance in order to avoid the strict requirements of the GDPR. Without anonymisation and pseudonymisation, it would in some contexts not be possible to provide the required data.

According to Article 4 No. 5 GDPR, pseudonymisation is the processing of personal data in such a manner that the personal data can no longer be attributed to a specific data subject without the use of additional information, provided that such additional information is kept separately and is subject to technical and organisational measures to ensure that the personal data are not attributed to an identified or identifiable natural person. The basic prerequisite for pseudonymisation within the meaning of the GDPR is that the identifiable data record is divided into a pseudonymised data record and further information [93]. This usually involves replacing the identifiable information with a pseudonym [75,93,94,95,96]. For example, the name is replaced by a number [75,93,94,95,96]. It must be ensured that the data subject can only be identified by using this additional information [97].

This is in contrast to anonymisation. The GDPR does not define this process. Only Recital 26 of the GDPR contains a corresponding definition of anonymised data. Anonymised data is therefore understood as information that does not relate to an identified or identifiable natural person, or personal data that has been anonymised in such a way that the data subject cannot or can no longer be identified. Therefore, in the literature, anonymisation is considered to have occurred when the personal reference is removed in such a way that re-identification is practically impossible [97,98]. Re-identification is no longer practically feasible if re-establishing the link between the person and the data would require a disproportionate amount of time, cost and effort [97,98]. For instance, anonymisation can be achieved by aggregating data sets, blending individual data points into larger groups to obscure individual identities [99]. Another example is data masking, where identifying information is replaced or modified to prevent identification [99].

However, this legal black-and-white understanding of data anonymisation does not always reflect reality. In real life, there are varying degrees of anonymisation and probabilistic chances of re-identification. If the likelihood of re-identification is very low, the relevant data can be considered anonymised.

3.3. The New Requirements for Data Protection Under the DGA

The EU has set itself the objective of encouraging more data owner to provide data free of charge through data altruistic organisations [100]. Data altruistic organisations are expected to ensure, due to their recognised status, that data holders trust these organisations, that the data serves general purposes, and are therefore willing to donate their data, Recital 46 of the DGA [101]. This is intended to increase the availability of data and thus enable easier access to this data.

However, registration as a data altruistic organisation does not come with minimal obligations; instead, further requirements are established for these organisations by the DGA [100]. The advantage, according to Article 17 (2) DGA, is merely that they can designate themselves as recognised data altruistic organisations within the European Union, be listed in a register, and use a corresponding logo [100,102].

3.3.1. The Intersection of GDPR and DGA

A significant issue arises when the data protection requirements of the GDPR intersect with the provisions of the DGA [100,103]. Article 1 (3) DGA stipulates that Union and national provisions on the protection of personal data are not affected [104]. Furthermore, in the event of a conflict, the GDPR regulations take precedence over the DGA provisions. Recital 4 of the DGA also clarifies that the DGA does not create its own legal basis [100,101,104]. Consequently, the provisions of Articles 6 and 9 of the GDPR are generally applicable [100]. In accordance with Article 2 (16) of the DGA, processing is only permissible on the basis of permission, rather than statutory exceptions, if the intention is to disseminate the data within a data altruistic organisation [100,101,105]. Therefore, data intermediation services must comply with data protection regulations [106].

The DGA introduces a number of considerations regarding the formulation of consent for the processing of personal data. In accordance with Recital 33 of the GDPR, non-specific consent for purposes of general interest is inadmissible [100,107]. At the time of consent, data donors are unaware of the specific purposes for which their data will be used [108]. The concepts of “broad consent” and “dynamic consent” may offer a potential solution, provided that their legality is definitively clarified [100,101,109,110]. Furthermore, the DGA imposes more stringent requirements than the GDPR with regard to the formulation of consent for the processing of personal data. Article 21 (3) DGA requires data altruistic organisations to provide a tool that allows data donors to easily withdraw their consent for data processing [100]. In the event that consent is withdrawn, the data altruistic organisation is unable to rely on another basis for processing in accordance with Article 2 (16) DGA [100,105].

Even with regard to the creation of the European consent form, this tension cannot be easily resolved. The European consent form is to be created in accordance with Article 25 of the DGA, following consultation with the European Data Protection Board and consideration of the recommendations of other stakeholders. However, this process does not guarantee that the form will comply with the data protection requirements of the GDPR. Furthermore, the use of the European consent form does not imply that the consent is lawful [111]. The review of the lawfulness of data protection remains the responsibility of the data protection authorities and the competent courts [100].

3.3.2. The Protection of Non-Personal Data

In comparison to the stringent requirements for processing personal data, the EU legislator imposes relatively few obligations on the processing of non-personal data. In this regard, the focal point is the Directive (EU) 2016/943, which establishes provisions for the protection of trade secrets. This includes data that is not held by public entities and often contains sensitive business information protected by confidentiality obligations [106]. Additionally, with the enactment of the DGA, further regulations come into play concerning the processing of non-personal data. However, these regulations primarily represent access restrictions to non-personal data.

Directive (EU) 2016/943 sets the framework for Member States to adopt rules to protect access to trade secrets [112]. In this respect, the Directive only aims to provide protection against anyone who infringes the sphere of secrecy established by the owner of the secret by unlawfully acquiring, using or disclosing trade secrets [113]. Trade secrets within the meaning of Article 2 No. 1 of Directive (EU) 2016/943 is information if it is secret, i.e., the information is neither generally known nor readily accessible, it has commercial value and is protected by appropriate confidentiality measures. In this respect, Directive (EU) 2016/943 restricts access to non-personal data to a certain extent, but does not create any direct barriers to processing.

Additional protection regulations for non-personal data also arise from the Data Act. The primary aim of the Data Act is to establish a legal framework that ensures fair access to data and its use in the digital economy [104]. While the Data Act primarily regulates the provision of data, it also includes the introduction of protective measures against unauthorised third-party access to non-personal data, which intermediaries must adhere to [104]. According to Article 2 No. 4 Data Act non-personal data is defined as all data that is not personal data. The definition of personal data is derived from Article 4 No. 1 GDPR, as cited in Article 2 No. 3 Data Act. If data, whether personal or non-personal, is made available to third parties by the user in accordance with Article 5 Data Act, the third party may only use the data for the agreed purpose and must delete it afterward, as stipulated in Article 6 (1) Data Act [104,114]. In accordance with Article 2 No. 12 Data Act, the term “user” is defined as a natural or legal person who owns a connected product or who has been granted rights to use the connected product on a temporary basis by contract or who receives services associated with the product [114]. Furthermore, Article 6 (2) Data Act delineates specific stipulations, such as the prohibition of utilising the data for profiling unless the user consents or disseminating the data to additional third parties [104,114]. This implies that the Data Act also delineates supplementary prerequisites for the transfer of non-personal data when users transfer data from manufacturers to third parties.

The safeguards introduced by the DGA go beyond the scope of the Directive (EU) 2016/943. According to the DGA, the processing of non-personal data, as described in Article 2 DGA, requires the unambiguous permission of the data owner [105]. This data owner, as clarified in Article 2 (6) DGA, may be a legal entity, including public bodies and international organisations, or an individual who, in relation to the data concerned, does not fall within the definition of an interested party. An interested party is authorised, in accordance with the applicable Union or national law, to grant access to certain personal or non-personal data or to disclose such data.

The concept of permission, which is analogous to consent in the context of personal data processing, signifies the conferral of processing rights [105]. However, the DGA refrains from providing explicit guidelines or criteria for obtaining permission [105]. Only Article 21 DGA articulates the requirements for data controllers, describing how they should transparently communicate information about the processing [105]. Ongoing discussions in certain contexts suggest the application of a standard comparable to that of consent for the protection of personal data [105]. The consent form explained in Article 25 DGA has been put forward as a recommended template for this purpose [101]. Nevertheless, freedom of contract applies within the scope of the permission [105]. In this respect, the DGA also justifies an access restriction for non-personal data, which, however, sets lower requirements than the GDPR for personal data.

3.4. Summary

Traffic management systems rely on an extensive databases for precise decision-making and forecasting, ranging from passenger information to the location data of transportation vehicles. Due to the potential personal relevance of this data, it is essential to carefully examine whether the GDPR applies.

The effective handling of the required data volume largely depends on anonymisation and pseudonymisation. These methods are crucial because accessing non-personal data is easier than accessing personal data. Anonymisation and pseudonymisation are not only legal necessities but also effective tools to ensure adequate data protection while enabling the accessibility of required data.

When data is to be accessed by data altruistic organisations, it is crucial to consider that these entities are subject to a range of additional requirements. The interplay between the DGA and the GDPR continues to present numerous challenges, including whether a suitable European declaration of consent could resolve this relationship.

4. Conclusions

This paper’s technical and legal analyses focus on the role of data in transforming the European Union’s transportation system, aligning with the objectives of the EGD. From a technical perspective, achieving efficient multimodal and intermodal coordination in transportation requires data sharing between stakeholders such as transport operators, infrastructure providers, and passengers. This data exchange is needed for improving synchronisation between modes, optimising routes, and meeting the 4 h door-to-door goal. Advanced coordination methodologies, capable of addressing the different dynamics of the system, are essential for managing the complexity of coordination and system optimisation. The approach integrates the strategic, tactical, and operational layers with distinct coordination methods, making it possible to achieve a combined optimisation, in which the focus on a passenger centric or a system centric approach can be followed.

One significant challenge is ensuring the availability of both personal and non-personal data in the required quality and quantity. In Europe, the availability of this data is largely shaped by the European data strategy, with the Data Governance Act (DGA) playing a decisive role in regulating access to such data. In addition to the consent requirements established by the General Data Protection Regulation (GDPR), the DGA introduces new permissions for using non-personal data. The DGA’s consent requirements are less restrictive than those under the GDPR for personal data, which makes pseudonymisation and anonymisation critical techniques to ensure data is available in sufficient quantities without violating privacy regulations.

From a legal perspective, compliance with both the GDPR and the DGA is paramount when collecting, processing, and sharing data. The GDPR imposes strict limitations on the use of personal data, requiring clear legal bases for processing, such as consent or contractual necessity. Meanwhile, the DGA encourages responsible sharing of non-personal data, including fostering data altruism. Anonymisation and pseudonymisation techniques are key to reconciling these legal requirements, as they ensure personal data is sufficiently de-identified to minimise privacy risks while enabling its use for transportation optimisation.

Handling data appropriately involves a careful balance between utility and privacy. For example, in coordinating multimodal transport, sharing passenger location data could help optimise public transport schedules and reduce delays. However, to comply with the GDPR, this location data should be anonymised or pseudonymised so that individual passengers cannot be easily identified. If pseudonymisation is used, the identifiable information (e.g., a passenger’s name) should be stored separately and securely, ensuring that it cannot be linked to the location data unless absolutely necessary and under strict conditions.

Further research includes identifying a viable combination of coordination methods, integrating them into a simulation framework, and evaluating the impact of each delay on stakeholder key performance indicators (KPIs). Given that the influence of delays depends on the specific transportation network, a generic simulation framework is needed. This framework will enable the simulation of specific use cases, allowing for the identification and evaluation of delays that contribute most to overall disruptions in transportation coordination. Further work will also have to include an analysis of the business network, considering available resources and activities from stakeholders, to evaluate the information and determine the associated costs. This will allow for a more accurate assessment of the value of information.

In conclusion, while technical innovations may greatly enhance transportation coordination, their success depends on the availability of the right high quality data and on carefully navigating the legal landscape. A balanced approach is required, where advanced data driven solutions are integrated with legal strategies to ensure both system efficiency and compliance with EU data protection laws. This approach is therefore necessary to realise a sustainable and efficient transportation system, in accordance with legal regulations.