/PILL: Physical Internet Living Lab

PILL: Physical Internet Living Lab

This page is also available in Dutch.

1. Intro

PILL (Physical Internet Living Lab) is a 3-year Flemish strategic fundamental research project (cSBO) lead by a consortium of imec, VUB and VIL. The project’s objective is to bring the current research on a Physical Internet for logistics and its principles into practice, laying the foundation of a general Physical Internet implementation framework for Europe and beyond.

The outcome is a blueprint for a Physical internet, supported by a software stack offering the first prototype of a PI network and application. All PILL deliverables are open and publicly available.

If you are interested in the PI blueprint and software stack, please get in touch to receive a download:

Contact us

2. The Physical Internet: a vision for the future of logistics

2.1 What is the Physical Internet

The logistics landscape today is riddled with challenges that hinder the seamless movement of goods between locations. Disruptions in the transport networks, unreliable planning, and a lack of transparency and collaboration between the many actors all lead to reduced operational efficiency.

Main visual

The Physical Internet (PI/π) is a vision for the reorganisation of the logistics sector. It redefines logistics as a decentralized, universally interconnected network of nodes where goods flow freely across the network in the most optimum way possible, mirroring some of the principles of the Digital Internet.

Figure right: in the Physical internet, goods move freely across nodes

“The paradigm change proposed through the Physical Internet is that logistics should be rethought as a system like the Digital Internet where networks would be interconnected through a common operating framework easing the breakdown of transport and handling loads. The Physical Internet allows progressively integrating currently dedicated logistics networks into a universally interconnected system.” - Montreuil, Meller & Ballot, 2010

In acknowledgement of this need for a more efficient and resilient logistics system, the European Commission has tasked the European Technology Platform on Logistics (also known as ETP Alice) to develop a strategic roadmap for the Physical Internet by 2040.

Video 1: The ETP Alice roadmap for the PI

Accept marketing-cookies to view this content.
Cookie settings

2.2 State of PI

The ETP Alice roadmap has since inspired various European and regional research initiatives and projects, most of which are focused on either proving the socio-economic benefits of synchromodal operations or on researching the possibilities of automating vessels and infrastructure. The result is an at times fragmented landscape of projects with no holistic technological approach on how to design and build such an interconnected and synchromodal network of assets and stakeholders.

PILL seeks to bridge this gap by establishing a holistic technical framework into which the individual use cases can integrate. At its core, PILL envisions an open, decentralized framework for an interoperable and interconnected logistics network.

Venn diagram with the three domains (socio-economic, digital and physical), with PI at the intersection

The main challenges on the road towards PI.

3. The PILL blueprint

Our technical design or ‘PILL blueprint’ of an open Physical Internet network comprises of three design principles.

  1. Network transparency
    All actors need to have -and contribute to- a holistic view of the entire logistics network with all its nodes, capabilities and transport.
    This transparency allows for optimum routing of transport.
  2. Decentralised interoperability
    All actors are connected through a decentralised network, that enables trust. By following a set of standards, logistics processes on this network are fully interoperable, enabling automation across value chains.
  3. Holistic, self-steering system
    All operations need to be self-steering and behave in a holistic way that maximally utilizes the available bandwidth of the system.
    This makes the system as efficient as possible, while increasing resilience against disruptions

3.1 The Network State: bringing transparency to the logistics network

One of the central goals of the Physical Internet is to enable transparency across the logistics network. The fundamental building blocks therefore required (nodes, capabilities and movers) were first outlined by Montreuil et al in their whitepaper Towards a Physical Internet: Meeting the Global Logistics Sustainability Grand Challenge (2010).

The capabilities identified within PILL (so far)

The capabilities identified within PILL (so far)

PILL’s contribution has been to (1) transform these building blocks into a standard for sharing logistics locations, services and capacity on an open network, and (2) consolidate the data into a virtual 2D map representing the entire logistics network: the Network State. By providing a holistic view of the logistics network to all participants, the Network State increases the discoverability and use of new nodes, routes or services.

The main advantages of this shared Network State are:

Figure 4: Overview of a first PI network state
  • Optimized route planning through the automatic consideration of additional routing options. Route engines on the PI can now consider all possible routing options.
  • Increased resilience by enabling real-time responses to disruptions in the logistics network.
  • Strategic optimizations achieved through simulations of potential changes to the network (e.g., assessing the impact of adding an extra train route to the network).

Figure right: Overview of a first PI network state

3.2 The PI-Client: Enabling interoperability on a decentralised network

Once a transparent network state is created, logistics actors can start cooperating on this logistics network. Since trusted collaboration is the pilar of the Physical Internet, collaboration should take place on a decentralised network. This PI network is grounded in the following four cornerstones:

  • Open, decentralized network: Behaviour & Governance rules for of a network for decentralized information sharing & interoperability.
  • Standardised data model & processes: Data and process standards for logistics transport, expanding on the existing DCSA standard, rooted in the larger UN/CEFACT
  • Universal PI-client connector: Software component that connects parties to the network & orchestrates interoperability between stakeholders.
  • Routing engine & simulation model: Component responsible for the holistic calculation of the most optimum flow of goods, provider & mode agnostic.

The blueprint’s PI-client is the key driver of the decentralized logistics network.
First, it acts as a connector to the logistics PI network. When installed, it operates as a digital agent that links to all other PI-clients and shares information about the network state to and from each participant.

Second, the PI-client serves as a platform for 3rd party applications and services:

  1. It holds a marketplace for decentralised PI applications that enable supply chain processes or strategic optimisations.
  2. It ensures that the different applications on the PI-network can interoperate with each other to automate and optimize processes.
Relationship between the digital PI client and the physical network

Relationship between the digital PI-client and the physical network

The PI client as a platform for 3rd party applications & services

The PI client as a platform for 3rd party applications & services

3.3 PI Simulation: a holistic, self-steering system

To design and comprehend the essential attributes of a Physical Internet (PI) system, we have developed a simulation model. This tool acts as a safe, risk-free environment for experimenting with and validating various components of the PI system. Namely, route planning and booking, and resilience against disruptions.

Our simulation tool follows the agent-based modelling methodology, where independent entities are represented as agents. These agents are designed to interact dynamically with their surroundings, striving to achieve specific objectives. In this context, an agent symbolizes a transport asset, a logistics facility, and a logistics operator utilizing the logistics ecosystem.

Figure 7: Agent based simulation model of a PI-network

By leveraging historical data from our project partners in simulation experiments, we were able to assess the advantages of transitioning container flows to a PI system, particularly in terms of emission reductions and enhanced responsiveness to disruptions. Additionally, the simulation model can function as a strategic digital twin of the PI system. By inputting the real-time network state, users can evaluate the potential impact and efficiency of future container movements or changes to the network state.

Figure right: Agent-based simulation model of a PI network

4. Validation: building & testing the PI blueprint

To validate our vision of the PI-blueprint, three software components were developed that represented the first functional prototype of a decentralised PI-network:

  • PI-Client: connects to the network, ensures interoperability with peers, and manages third-party applications.
  • Network State Platform: provides visibility to the network and its data.
  • PI simulation: simulates container flows over the proposed physical internet, producing detailed measurements
The first PI network prototype: three software components

The first PI network prototype: three software components

After developing and integrating these components into the proof-of-concept demo, a two-phase validation track was established to evaluate the effectiveness of the blueprint in terms of:

  • Improving the resilience of the network by evaluating its ability to withstand disruptions and find alternative solutions.
  • Enabling interoperability between stakeholders and processes on a decentralized network;

4.1 Risk-free environment validation

The PI simulation tool, played a pivotal role in verifying the functionality and efficiency of the blueprint's key components (booking & routing) at different scales. By simulating these processes on a greater scale, the tool provided invaluable insights into the operational feasibility and potential bottlenecks of the PI system, enabling us to built a scalable system.

Through simulation scenarios, comparing PI with the “Business as Usual” scenario, we quantified the improvements that the PI system could offer. By simulating both scenarios and analyzing the data, it was possible to draw clear comparisons in terms of efficiency, cost-effectiveness, and environmental impact. Our simulation experiments proved that a Physical internet, following our blueprint, will optimise the flow of logistics in a way that significantly reduces emissions and improves asset utilisation. You can find a detailed view of our results in our final report (Sept 2024).

4.2 The PI-Client Living Lab

The PI-Client Living Lab consisted of a 2-week field test of the PI-client and the decentralised route planner. For this test, ten participants (forwarders, transporters and terminal operators) who operate on the corridor of the Albert Canal (Antwerp) installed the PI-client and used the route planner to create a network state. For the next two weeks, forwarders used the route planner to find and book available routes on the PI-network.

Accept marketing-cookies to view this content.
Cookie settings

It was through this Living Lab test that the world’s first successful booking for a container on the Physical Internet was made.

5. Implementation of PI: integrating the PI client in data spaces

With the PI-client, PILL has demonstrated the feasibility of interoperability on a decentralised network. To scale the Physical Internet to a commercial application, the PI-client must evolve into a plug-and-play solution adept at addressing all potential concerns about joining in and collaborating on a decentralised network. However, the current technology of creating a decentral, interoperable network still faces several scalability challenges, such as discoverability, identification and agreements.

These challenges align closely to the concerns being researched in the domain of “Data Spaces”, an emerging technology ecosystem centered on data sharing through decentralised networks. The similarities between data spaces and PI present an opportunity to combine both technologies and leverage the benefits of both Data Spaces and the Physical Internet.

The principles and concerns of Data Spaces

The principles and concerns of Data Spaces align closely with those of a PI network

The final phase of the PILL project will therefore focus on setting up a third test, aimed at integrating the PI-client into the data space architecture. This merger of concerns and solutions from data spaces into the PI blueprint is expected to lead to the creation of a logistics PI data space that solves several scalability challenges currently hindering the PI-client.

6. Future projects: beyond PILL

PILL has already made significant strides in advancing the ALICE roadmap and the State of the Art (SOTA) on the PI by:

  • Building a framework for improved transparency in logistics nodes and services through the definition of a network state;
  • Developing a decentralized network via a PI-client that enables interoperability across all nodes in the network;
  • Addressing discoverability on a decentralized network by leveraging the foundations of data spaces.

By taking the first steps in integrating the PI in data spaces, PILL has established a first scalable framework and demonstrator of a Physical Internet. However, a key question remains unanswered.

Data spaces centre around data sharing technology. The primary concern in logistics, however, is ensuring the interoperability of processes and operations between participants. In many cases, data sharing is but a byproduct of the more crucial process sharing.

Creating a true PI data space therefore requires expanding the scope of data spaces to cover process sharing. As such, the PILL team is currently working on a follow-up project to build a process sharing framework for data spaces.

For deeper insights into process sharing, refer to this paper by research institute TNO.

Process sharing: the framework for a PI Data Space

Process sharing: the framework for a PI Data Space

If you are interested in participating in our ongoing research, please reach out