Mobility as a Service (MaaS) promises to dissolve the boundaries of today's transport network through the integration of transport modes. However, stakeholders such as the operators, public transport authorities, or end-user routing application developers, are facing interoperability barriers in the way of providing a seamless travel experience. To identify, understand, and devise a plan for integration, we must first be able to clearly define the interoperability requirements of MaaS. In this paper, we propose a framework consisting of 6 layers: (i) the Stakeholder layer, (ii) the Data layer, (iii) the Infrastructure layer, (iv) the Ownership layer, (v) the Auth layer, and (vi) the Connection layer. In each layer, we outline the opportunities for Semantic Web technologies. In addition, each layer is equipped with use-cases to illustrate the application of the framework. Through the Linked MaaS framework, we formulate a baseline that extends the research focus of semantics in transportation and builds a guideline for the development of an explicitly defined and understood MaaS.
Mobility as a Service (MaaS), an emerging concept that promises to dissolve the boundaries of today’s transport network, continues to face a manifold of closed doors. The progress towards the seamless vision of MaaS, which provides users with a carefree experience of planning, booking, and payment of transport services, requires integration at multiple levels of the MaaS ecosystem, including data integration for journey planning, ticketing and payment integration, and the integration and alignment of stakeholder’s commercial goals. To identify, understand, and devise a plan for integration, we must first be able to clearly define “What is a MaaS ecosystem and what are its interoperability requirements?” Copyright © 2021 for this paper by its authors. Use permitted under Creative Commons License Attribution 4.0 International (CC BY 4.0).
It can be argued that having a clearer vision of what MaaS and its
interoperability requirements are is not a significant solution to the barriers facing
MaaS. Yet, the lack of a common understanding of MaaS, its potential, and an
agreement over its fundamentals leads to an unresolved business model which, in
turn, leads to a lack of interest in stakeholders contributing to the development
of MaaS. In pursuit of finding the answer to this question, a thorough literature
review was used in an attempt to provide an unclouded understanding of MaaS.
However, there exists over 15 definitions attributed to ”Mobility as a Service”
in the literature with no agreed-upon definition [
The paper begins by presenting the background behind the study and a summary of the related work. Subsequently, an overview of the Linked MaaS framework is presented which is then broken down to its constituent layers. Each layer is explained thoroughly demonstrating its applicability through example use-cases. The paper is then concluded with the vision for the Linked MaaS framework and a few examples of synergistic use-cases that benefit from the framework holistically. 2
When MaaS was first coined by Hietanen [
Meadows [
A notable framework, by Sochor et al. [
To compose a solution that harmonizes the semantics of MaaS, we break down the MaaS ecosystem into its constituent sub-systems. Each sub-system is illustrated as a layer of categorized MaaS elements and their interconnections, with a distinct function for each layer. Together, these layers constitute the Linked MaaS framework as summarized in Figure 1. This framework is specific to MaaS, and not transport in general, in the sense that it focuses on enabling transport integration through MaaS. The framework consists of 3 Graph layers – comprised of both nodes and relationships: 1) The Stakeholder Layer, 2) The Data Layer, 3) The Infrastructure Layer,
The Graph layers can be viewed as an independent domain of analysis. In addition, we define 3 Linking layers – comprised of relationships only: 4) The Ownership Layer - Linking between Stakeholders and Infrastructure 5) The Auth Layer - Linking between Stakeholders and Data 6) The Connection Layer - Linking between Data and Infrastructure The proposed Linked MaaS framework introduces 6 interlinked layers, that each can benefit from using Semantic Web Technologies (SWT) to model, analyse, and formulate solutions for issues within each layer. Throughout the following subsections, we will discuss the functions of each layer and the reasoning behind its development. Furthermore, we will discuss the recommended Semantic Web technologies for diferent use-cases in each layer, in addition to, other supplementary technologies. 3.1
The first layer of the Linked MaaS framework is The Stakeholder Layer. As
the name suggests, this layer consists of the stakeholders of MaaS as the
elements of focus, whilst also modelling the interconnections between them. In
MaaS workflows, it is often the case that a single decision can have direct or
indirect efects on any MaaS stakeholder, whether taken collectively or
individually. In such a case, delineating the relationships and dependencies between
these key players is needed. The main function of this layer is to build an
abstract model of the business ecosystem of MaaS. Kamargianni and Matyas [
While business agreements and contracts will heavily depend on and benefit from this layer, it was rather included in the synergistic use cases as such contracts will include agreements on data sharing and access, risk distribution, and various factors supplemented by the other layers in the framework. 4 https://www.w3.org/2005/Incubator/decision/XGR-decision-20120417/
Sample_Decision_Ontology.html
A main area of concern within emerging transportation technologies is the
representation of data, syntactically and semantically. The digitization of
transportation data led to the development of multiple standards such as GTFS static5,
GTFS realtime6, GBFS7, MDS8, The TOMP-API 9, NeTEx10, SIRI11, DATEX
II12, OSLO Mobility13, and more. When a standard is developed, it is common to
define it up to a certain level of granularity, leaving certain issues open [
Data, in the MaaS ecosystem, is heterogeneous and present in silos. It is not
equipped with the metadata necessary to enable data interoperability. In
addition to multiple standards, there still exists a vast amount of data that does not
even follow a standard [
Some of the use-cases envisioned for the data layer are listed below: 5 https://developers.google.com/transit/gtfs 6 https://developers.google.com/transit/gtfs-realtime 7 https://github.com/NABSA/gbfs/blob/v2.2/gbfs.md 8 https://github.com/openmobilityfoundation/mobility-data-specification 9 https://github.com/TOMP-WG/TOMP-API 10 http://netex-cen.eu/ 11 https://www.vdv.de/siri.aspx 12 https://www.datex2.eu/ 13 https://data.vlaanderen.be/ns/mobiliteit – Automated integration of data from heterogeneous sources through the alignment with existing standards. – Performing simple and complex queries and algorithms over MaaS data (e.g. querying for availability of bike lanes, running route-optimization algorithms). – Reasoning over the graph to derive new information and conclusions (e.g.
Inferring which factors afect the operations of which modes, for example, an if rule can state that if an increase in a specific factor (e.g trafic) afects a property of a transport mode (e.g decreases the speed) then this factor should be taken into account by the routing algorithm ). 3.3
There is a distinction between servitizing transportation and servitizing other industries, of which some have witnessed successful models disrupting preceding distribution models. The implementation of MaaS relies on the existence of physical infrastructure. Physical transportation infrastructure is a prerequisite for actualizing MaaS. How can we shift people to more sustainable modes of transport if we are lacking the infrastructure for it? Promoting bikesharing, for example, requires bike lanes for a safer commute. Shifting to electric cars requires charging stations. Major investments in public transport is needed to set up railways and subways. The significance of infrastructure to MaaS necessitated its inclusion as the third layer of the Linked MaaS framework. Considering the technological advancements in transport, the core elements of the infrastructure layer include both physical and digital infrastructure. Digital infrastructure refers to the ticketing infrastructure, routing algorithms, MaaS aggregation platforms, cloud infrastructure, and others. The main function of this layer is to provide a digital twin of both physical and digital infrastructure as well as the relationships between them. As this layer connects to the stakeholder layer, it will represent all the assets owned by the stakeholders. Independently, the layer can be used to investigate the following scenarios: – Through graph theory approaches, various analysis can be performed to determine System Resilience, identifying the weakest links and bottlenecks in the infrastructure. This can aid in deriving multimodal disaster mitigation and recovery plans. – Analysis of the system to identify missing infrastructure links at diferent locations, both physical and digital, which require installation. Through such analytics, investments can be directed to the right infrastructure sectors and locations. – Inter-dependencies between modes and the dynamics of a multi-modal system can be derived from the model for usage in smart city mapping and planning.
After discussing Graph layers 1, 2, and 3, we now look at forming interlinks between these layers. The Linking layers represent relationships between elements from diferent graph layers. These layers are discussed in the subsequent sections. 3.4
There are relationships between MaaS stakeholders and the Infrastructure layer. These relations can include which infrastructure belongs to a stakeholder’s assets, who is responsible for the maintenance or development of specific infrastructure, etc. These questions mainly depend on the ownership of assets. Therefore, the main function of this layer is to model the links between the stakeholders and the infrastructure. A few use-cases envisioned for this linking layer are: – Analysis of risk. Through the links between the stakeholders and their assets, an optimum risk distribution model can be determined for a given MaaS ecosystem. – Maintenance schedules and work orders can be linked taking into account the efects of surrounding infrastructure. – The layer can be used in synergy with other layers for the development of suitable contracts and agreements based on stakeholders’ assets and liabilities. 3.5
A major debate within MaaS falls under the umbrella of data management. Some of the questions include: – Who will have the authority over MaaS users data? – Who is responsible for data maintenance and quality? – Who is authorized to access a specific type of data? – How do we authenticate the user/agent accessing the data? – Will MaaS Providers contribute to the monopolization of data or will there be a decentralised data access? To resolve this debate, we propose the Auth layer. The main function of this layer is to model the links between stakeholders and data. The incorporation of the Auth layer in Linked MaaS can aid in the following use-cases: – Explicit identification of read-write access rules to diferent data entities, perhaps through the Access Control List Ontology. – Explicit definitions of roles and responsibilities of the stakeholders for data management. – Examining the efects of diferent data access distributions on the stakeholders of the ecosystem (e.g. business opportunities through Open Data). – Incorporation of data pods following the Solid14 Project for decentralization of user data.
– Enforcing data policies through validation rules. 3.6
Although the Data layer and the Infrastructure layer appear to overlap, there is a clear distinction between them. The Infrastructure layer models the physical and digital assets of the ecosystem, whereas a part of the Data layer would consist of data derived from that infrastructure. For instance, the infrastructure layer would model railway tracks whereas data from sensors attached to these tracks are modelled in the data layer. Between these two layers, we propose a relationship layer, The Connection Layer, where the main function of the layer is to model the interconnections, policies, and business rules between data and infrastructure. Example use-cases envisioned for this layer are: – Can be used to power the development of Internet of Things in transportation, providing insights on which parts of the infrastructure is connected over the internet, what types of connections do they have, for instance, having vehicle location data for a train but lacking sensor data for facility management. – Identifying physical entities from which data is generated can provide insights on connectivity gaps. For example, identifying stations and infrastructure for which schedules are available can help identify which infrastructure is lacking connectivity to the routing algorithms. – Policies on the usage of data for certain infrastructure can be formalized within this layer. 4
In this paper, we presented the Linked MaaS framework and discussed 3 Graph Layers and 3 Linking layers, justifying their function in the ecosystem. In alignment with the results of the literature, the layers contain all categories identified 14 https://solidproject.org/ which were (1) Business, Policy, and Stakeholders, (2) Data and Technology, and (3) Infrastructure, noting that the inclusion of policy concepts is distributed throughout the layers rather than constituting a layer of its own. We envision the development of Linked MaaS to be a catalyst for actualising MaaS. As part of an ongoing research, we are focusing on works that will contribute to the Stakeholder layer, the Auth layer, and the Data layer. We invite researchers of MaaS and the Semantic Web community to bring the Linked MaaS vision to reality through novel contributions to the 6 layers.
Under the previous subsections of Layers 1 to 6, we provided examples of usecases that the individual layers are expected to handle as part of their function. From a holistic perspective of Linked MaaS, there is an opportunity for synergistic use-cases which rely on the collective use of the layers. Some examples are: – Generation of data-driven contracts, based on the analysis of the stakeholders’ relationships with infrastructure, with data, and with each other. – A simulation engine can be derived to predict the efects of adding, removing, or changing any elements in the system (e.g. a merger between two companies, construction of a new highway, incorporation of a new technology). – Data Governance, including policies, roles, standards, and metrics, can be formalized through the Stakeholder layer, the Auth layer, and the Data layer.
Semantic Web technologies were chosen as the core tools powering the Linked
MaaS framework for their extensive capabilities to link data from heterogeneous
sources, develop abstract models that reflect the complexity of the world in the
form of simpler ideas, and reason over encoded knowledge to infer new and
meaningful information. The high-level concepts for all the layers, likely
represented as ontologies or vocabularies, can serve as a foundation upon which a
Linked MaaS knowledge graph is constructed specific to given geographic
regions where MaaS operates. This will set a roadmap for MaaS to achieve an
Adaptive level of interoperability - highest level in the Enterprise
Interoperability Model [