LINC project Journal 4 - get an update about the Albertslund project

The Danish legislative process for approval to operate self-driving vehicles on public streets is one of the most restrictive in the EU, and includes many parallel approval steps involving the Danish Transport Authority, The Danish Road Directorate, and a third party safety assessor. In the LINC project the application, including evaluations from the third party assessor, was sent to the Danish Road Directorate in June 2020. Several additional questions from the authorities had to be answered, and the official processing of the application could not start until November 2020. 

After the authority's review of the application, a four weeks long public hearing period, and a processing period, the Danish minister of transport declared the test with self-driving shuttles on DTU Lyngby Campus approved on March 18th, 2021. This was the starting point of an intense period for the LINC, including education of test drivers, mapping of the route with the vehicles, and test driving. The final approval to start operation with passengers on board at DTU Lyngby Campus was made by the assessor on April 12th. 

In total, the LINC project has spent around 30 calendar months to get permission to run the self-driving shuttles on public streets on SAE level 3 (i.e. with a safety driver on-board). After a long struggle, this important milestone was achieved by the project, and on April 19th the rubber could hit the asphalt, with the passengers on board, for the first time.

The DTU Lyngby Campus test is more than self-driving shuttles. It also includes an intelligent “Smartphone sensing platform” consisting of Bluetooth beacons that sense the passengers’ smartphones, an infrastructure that can combine various data sources, and a backend system for data analysis. 

Valentino Servici, DTU, explains that the smartphone sensing platform can be used for multiple purposes: 

• It provides accurate positioning of passengers and recordings of anonymous movement patterns, data that can be used to tune the operation to respond to the actual demand.
• In particular, the sensing platform determines how many passengers there is onboard a vehicle. This information can then be used in the real-time operation, to determine if the vehicle can pick up more passengers.
• The platform enables smartphone data to be combined with e.g. weather forecast data and vehicle data, combinations that can be used for control of the shuttles in real-time.
• It can push information or surveys to users, triggered by events. For example, passengers that were on board during harsh braking can be asked about their experience of that directly after the event.
• The platform has a dashboard for operators, which shows real-time metrics, and also allows for analysis of operational data for service improvement. 

The smartphone sensing platform has been developed during the last year and was tested together with the shuttles on closed roads on several field tests. The latest field tests took place on February 17th, 2021, and March 22nd, 2021.

During the fall and the winter, the project members have worked intensively on updating the test design to respond to changes in the test caused by the delay of the legal approval and by Covid-19 restrictions. The fact that the legal approval process has been very long and uncertain has led to replanning, both in terms of when to run the tests, and what to include. In addition, the Covid-19 pandemic has introduced even more uncertainties due to lockdowns and limitations in people's movements. The final test design was set at the end of March 2021 and includes an operation with passengers at DTU Lyngby Campus from April 19th to July 2nd. The route for the test has been shortened and one bus stop removed due to road maintenance work at Niels Koppels Allé, the southernmost part of the route. The final route is shown in the Figure below. Unfortunately, even if the LINC  project continues beyond July 2nd the tests cannot, due to further road maintenance covering most of the route network.

To prepare for the tests, around 488 test passengers have been recruited. Due to the Covid-19 pandemic, DTU Lyngby Campus has been closed for several months. Currently, the campus is expected to open up to around 20% of normal activity during the LINC test period. Therefore, emphasis in the recruitment of test passengers has been on people that actually are at the campus site, and recruitment has been done physically at the Campus.
The relatively short test time, caused by the complex legal environment, and the restrictions due to Covid-19 have shifted the focus of the tests. Instead of testing an optimized demand-responsive mobility service, the focus will be on testing robot-human interactions, developed user services (like event-triggered surveys), and collecting user data for mathematical simulation models. By using simulation models, the project seeks to use the real transport demand data collected by the smartphone sensing platform, and explore the trade-off between offering a pre-scheduled service synchronized with the light-rail and reserving capacity on the shuttles for on-demand services. In this way, also the test will provide useful information about on-demand services to decision-makers.
To achieve the results above, the project is dependent to have an approval of a request for project prolongation for three months. This will allow the project to gain more operational time and also the chance of having an AV operation at DTU Lyngby Campus in a post-Covid-19 reopening of the campus.  

The long-term goal for the LINC project is to contribute with knowledge about how smart mobility can be a part of sustainable urban planning. To fulfill that goal the LINC project has performed several activities during the last year:

• Workshops with urban planners and decision-makers have been executed and more are planned for the future. 
• Simulations to explore the potential impacts of shuttles as a complement to the new light rail in Hertsted Business park have been set up. 
• The project has initiated a study together with ViaTrafik, to explore how first/last mile self-driving shuttles as a complement to the Copenhagen Light Rail can be integrated into urban plans. In particular, the two urban transition areas Hersted Industrial Site and Gladsaxe Erhvervsby will be studied. The aim of the study is to support sustainable urban planning processes. Specifically, the study will investigate the needs of new physical and digital infrastructure for this type of service.

While waiting for the start of the operation, simulations have become an important tool. Professor Jeppe Rich, DTU, explains:

We want to look at the first/last mile service in conjunction with the light rail. The simulation will primarily be on Hertsted Industry Park and is based on origin-destination matrices that show how people move.

Professor Rich explains that, for example, their model aims to answer how many buses will be needed to meet a certain demand. They are also performing a cost-benefit analysis to explore the effects of the service on a societal level.

The LINC project has identified the long legislative process to get the approval for the operation of the shuttles as one main barrier for the development of innovative mobility solutions in Denmark. Therefore, the project has done several efforts to change the legislative landscape related to new mobility in Denmark: A Letter of recommendations was sent to the Danish minister of transport, six recommendations for improving the Danish pilot scheme for automated vehicles has been identified and communicated, and a press release based on a comparison of the Swedish and Danish processes has been published. In addition, project members have been engaged in a national network for projects with self-driving shuttles. This network has provided a platform for the exchange of experiences but also formed a stronger body concerning legislative questions.

The global Covid-19 pandemic and the related lockdowns and restrictions have impacted the project in multiple ways. 

The lockdowns of Denmark have led to that fewer people than normal will be at the DTU Lyngby Campus during the test period. This limits the number of potential users and changes the movement patterns of those that are at the campus. The pandemic restrictions have also imposed a limit of max 2 passengers on the bus, which also impacts the actual use of the busses.  One project member states that 

The implication can be huge - without people walking around and taking the bus we don’t have any data.

The project has replanned the tests to focus on what can be tested given the circumstances: the technology and the user’s adoption of the technology, rather than the full service that originally was the plan. Furthermore,  the project requested a project prolongation for an additional three months. This will give the project enough time to finalize the processing of the data from the three months testing period ending July 2nd, 2021. It could also provide opportunities for additional operational time if needed if circumstances allow (given road maintenance work, financial resources, etc).  

Besides the implications for the actual tests, Covid-19 has likely also caused delays in the approval process, since Covid-19 related questions have been prioritized by authorities. 

The pandemic has also slowed down the project's internal work in different ways. Important parts of the work were planned to be performed in collaboration with experts e.g. from Nobina Technology in Sweden and the vehicle supplier in France. The travel restrictions during the pandemic have put all such collaborations on hold.  Furthermore, The LINC project is a truly cross-disciplinary project, where collaborative dialogues and co-creation to explore and mutually understand the different perspectives are key. Project members have experienced that having such dialogues online instead of in-person has been challenging. One project member explains:

It [Communication during lockdown] is like we are shouting to each other through a garden hose.

The challenges with digital co-creation meetings and workshops have put an extra burden on many of the project members.

There are two hard challenges in the project, Leadership for innovation and Communication with target beneficiaries. At the core of these two challenges is the fact that the actual tests with the self-driving shuttles are significantly delayed due to external reasons (very long legislative process and COVID-19). The dialogue with decision-makers and target beneficiaries is dependent on if the project can provide tangible evidence or only can talk about the potential benefits. 

On the national level, the project has already contributed to improvements by participating in the national dialogue on legislation for self-driving vehicles. 

On the regional and city levels, real evidence is central for the project to have a significant impact on urban planning. With the very short test period, that will occur during a time when restrictions due to Covid-19 limit mobility, it is still uncertain whether the project will actually gain the results that they need. The project has met this challenge in several ways: 

1. The project plans the tests so that as much data as possible can be collected during the limited time. Then, the researchers can analyze the data after the test has ended.
2. The tests are planned to collect data from what can actually be tested (e.g. the technology rather than the service).
3. Simulations are used instead, to explore the potential of self-driving shuttles in first/last mile services as a complement to the light rail. Simulations can compensate in some perspectives, and even provide more information as they can study areas that are not included in the real-life tests. However, user acceptance and adoption of new services can not be measured in the simulation.
4. Together with the two municipalities in the project an “ideation catalog” presenting ideas that have arisen during workshops in the project. From this, the idea is to illustrate opportunities for the future. This is ongoing work, and a way to build leadership for innovation. The ideation catalog is supplemented by an analysis of the potential for Greater Copenhagen Light Rail linked to a first/last mile self-driving service for Hersted Industrial Site and self-driving shuttles impacts on Sustainable Urban Mobility Planning (SUMP) framework which eventually will lead to new approaches for SUMP.
5. The project is also performing a desktop exploration of sustainable business models for innovative mobility as a part of urban planning.
6. The project has increased its efforts in communication with for example Copenhagen's Light Rail board and other stakeholders. 

One main challenge in the project is to get politicians and decision-makers to “buy-in” on the ideas and plans from the project. Autonomous mobility is not a part of the current city master plans in the municipalities, but one success criterion for the LINC project is that it will be in the future. To take one major step towards this, a workshop, organized together with ViaTrafik, is planned for June 2nd, 2021. The workshop gathers key stakeholders including urban planners, traffic planners, smart city consultants, and the public transport agency Movia, to formulate a vision for connected, and shared automated mobility. Anchoring of the vision, and development of a strategy for the implementation of self-driving shuttles to support sustainable cities, is crucial for success. It is important for the LINC project to show the benefits or at least potentials for the decision-makers. The aim of the LINC project is also that one or several of the stakeholders at the workshop will take over the ownership of this vision for future development, or that a new project can be founded on this vision and strategy. The fact that the actual tests with the autonomous vehicles finally have started will hopefully raise energy levels in this question again.

 There are both technical and legal barriers that hinder the implementation of self-driving shuttles, or expressed by the project members: 

We have learnt that the technology is not ready yet to be implemented

Our understanding of the project has shifted - from a test of a new public transport service to a complex technology development project

The LINC project is not alone in this insight. For example, the heavy truck company Volvo recently announced that they would pause their flagship project due to legislative challenges. This was also discussed in the UIA expert journal no 3 for the project, referring to the Gartner hype cycle, where self-driving vehicles have gone from being at the “Peak of inflated expectations” to be in the “Trough of disillusionment”. That said, one other main learning is that it is through pushing the limits things can change.

Waiting and waiting for the legal approval has given the project members the feeling that the air has got out of the balloon. The delays have made people disillusioned about the technology and it is sometimes difficult to attract their attention again. But now when the tests finally can start, there is great external interest!

A resilient project is a project that does not fail even if it goes wrong (the notion of resilient projects was discussed in project Journal 3), and when designing and running complex innovation projects, resilience is a very important aspect. Both in terms of results from the project, but also for keeping the motivation and engagement among people engaged in the project. 

Looking at the LINC, it is obvious that the project has missed one of its original main goals: to run two six-month-long tests with self-driving shuttles and that include tests of an innovative on-demand service. At the same time, the project has achieved many other things: insights about the importance of harmonization of technology, legislation, and policy developments, the development of a technically advanced smartphone sensing platform for detailed mobility demand measurements, and first insights on how future mobility solutions like autonomous collective mobility can support sustainable urban planning. By acting agile and open to new methods (e.g. use of simulation models), the project has managed to adapt to the circumstances and made the most out of the available results.