CPS4EU, InSecTT and Arrowhead Tools: understanding interdisciplinarity in European projects
‘Interdisciplinarity’ is defined by the Oxford Dictionary as “the quality or fact of involving different areas of knowledge or study.” This concept is particularly relevant to the creation of modern Intelligent Digital Systems, which is simply not possible without the involvement of parties from across the value network and Europe as a whole. What does interdisciplinarity mean in a practical context? And how are European projects implementing it? Three Inside members share their views and experiences.
The dimensions of interdisciplinarity
Of course, interdisciplinarity can mean different things to different individuals – collaboration in academia, for instance, typically looks very different to cooperation between OEMs. Nonetheless, Jerker Delsing, Michael Karner and Philippe Gougeon see eye to eye on many of the nuances. “It’s kind of a funny word!” begins Jerker, Professor of Cyber-Physical Systems at Luleå University of Technology in Sweden. “If we are too far from each other, the journey becomes quite long. But with reasonable distance, there’s grounds for making something new. To me, interdisciplinarity means that when you meet people who have different views and backgrounds, your background can provide new light into their picture. To begin with, there might be one-way understanding. But if we can get a mutual understanding, things will really take off.”
“For me, interdisciplinarity essentially means looking beyond one’s own nose and leaving behind traditional ways of thinking that you may have inherited from your discipline,” says Michael Karner, Lead Researcher in Embedded Systems at Virtual Vehicle in Austria. “With communication across traditional borders, you’ll be able to achieve high-value results that would not have been possible by staying within your own discipline. The topic of interdisciplinary has many different dimensions but it often comes back to cooperation: accepting and understanding different perspectives, finding a common ground in communicating and developing a joint solution. If you think about it, it’s a bit like bringing together different electronic systems to form a system of systems. Of course, this could just be my bias as an engineer!”
“I agree with all these points already mentioned,” adds Philippe, Collaborative Project Director at Valeo Comfort and Driving Assistance. “An additional perspective is that one of the great benefits for a company like Valeo is working in large European projects and joining developments by companies or experts in other domains. They have a technical level which is already quite advanced. There is also a transversal dimension related to the technical level of microelectronics, hardware and software developers, system architects and people working in operations and recycling. There is a need to create a link between all these technical activities. This is what we’re targeting in my project and also in the way we develop new projects.”
Something like a company
For Philippe, the project in question is CPS4EU, which brings together large and small companies and research centres to strengthen Europe’s cyber-physical system value chain in terms of automotive, industry, energy distribution and use-cases for SMEs. As Valeo is currently moving into new mobility markets such as automated shuttles, robo-taxis and drones for delivery, CPS4EU focuses on the key enabling technologies of computing, connectivity, sensing and cooperative systems. In doing so, they’ve built up an ecosystem which is the full-time equivalent of 130 people for three years.
“It’s like a small company with all the different competencies and products that we will use to develop our complex cyber-physical systems,” explains Philippe. “We have small companies working on things like sensors or components to perform AI calculations. We also have the providers of the use-cases. Cyber-physical systems are quite complex, so we can no longer perform all-vertical development by ourselves. We need to work in a new network using pre-developed, pre-validated and pre-integrated subsystems. This is the purpose of our project. On top of that, we have mostly academics – but also some SMEs – working on tools to develop and orchestrate all of these new cyber-physical systems.”
As an example of interdisciplinarity within the project, Philippe is spoilt for choice between any of the 16 use-cases. “For example, there’s a lot of cameras on current vehicles and there will be even more in the future. With our colleagues from CEA, we’ve created a team to work on best-in-class deep neural algorithms to detect and classify objects in the surroundings of the car using pixel segmentation. Every pixel is classified as an object so that the algorithms can define their trajectory and apply steering, power or braking.” This is complemented by a computing architecture which has been optimised for performance and power consumption when running this neural network software. The result is an example of technology which spans the entire chain from hardware to software, without which the emerging domain of automated driving would not be possible.
Challenging initial assumptions
Like CPS4EU, Michael views the InSecTT project as something of a small company, thanks in part to the improved organisation of European projects in general over recent years. This includes continuity between projects, as InSecTT extends the wireless connectivity and interoperability successes of the DEWI and SCOTT projects with the dimensions of trust and Artificial Intelligence (AI). As the project manager and coordinator of the predecessors, it was only natural that Michael take on this role again in InSecTT.
“There are two things I’d like to highlight on interdisciplinarity. On one hand, we are specifically working on trustable systems. What does it mean for a user to have trust in a system based on AI?” asks Michael. “For critical systems or systems that are deeply embedded in our lives (like health systems), the user needs full trust or it won’t work. For this reason, InSecTT has psychologists working closely with the engineers to provide guidelines for the development of trustable AI systems. In the end, you need to make sure that the intended user of your AI-based IoT system understands how it makes decisions – that it’s not just a simple black box – and that the AI ‘understands’ what the user wants and needs. The second example of interdisciplinarity is an open innovation approach. We are doing open innovation within the consortium but, in the last phase of the project, we will also do it externally. Here, we are already seeing a great uptake of ideas from very different applications that the original developer did not even think of.”
One such example is a cost-effective wireless IoT switch, which was developed in SCOTT and further enhanced in InSecTT. This was later identified as a good fit for the aviation industry, such as in detecting if seatbelts are fastened and lifejackets are in place. The key partner, TU Delft, subsequently won a prize from Airbus and have since launched a start-up based on this idea, despite it not originally being in their plan. The project will continue to build on such successes and uncover previously unforeseen applications as it enters its final year.
The importance of openness
As these projects show, a core theme of interdisciplinarity is the application of various disciplines to solve practical issues. The same is true of Arrowhead Tools, which looks into how internet technologies can be used to create automation and digitalisation solutions. “What interests me a lot with this project,” says Jerker, “is that we’re working with real-world industrial and societal problems to which we can actually make a difference as a university. This is also gives me, my students and my colleagues concrete, realistic things to work with, not things that we dream up in an academic bubble.”
With a large mix of 54 companies and 36 RTOs or universities, the project combines the perspectives of end-users, suppliers and knowledge providers in order to experimentally verify and validate common technology in 21 use-cases. Given the wide scope, from building management to teaching, the need to break out of technological silos is self-evident. “We have one project with a common technology that we cannot divide, not many sub-projects that work on their own,” notes Jerker. “You can talk about interdisciplinarity in science and research, but here we see it in the overall technological landscape. And this is also part of something much larger: open source, which is a lot of the common technology we deliver in Arrowhead Tools and which requires quite different thinking compared to proprietary technologies. Many companies are not yet experienced in utilising and making money out of this. This is another part of interdisciplinarity: how can we break with the old ways of doing things?”
The idea is clearly catching on throughout Europe, as Jerker has already given four two-hour webinars to other projects and standardisation bodies in the spring of 2022. “A number of companies are now selling results to the market,” he continues. “Of course, they sometimes get questions that they cannot directly answer and they filter this back into the project and ecosystem.” In the production line creation of bathrooms, for instance, Arrowhead Tools encountered a partner which was struggling with the translation of data from one of their architect systems into robots that produce certain parts. With an automation approach developed in the project, they’ve reduced the time taken to produce such parts from seven weeks down to a few days. This is expected to fall to just 90 seconds by the end of Arrowhead Tools. In short, an interdisciplinary perspective – in this case, considering automation over the traditional manual approach – can allow companies to become much more competitive.
A tool for bigger things
As ECSEL-funded projects, CPS4EU, InSecTT and Arrowhead Tools form part of the long-term goals of both Inside and the European Commission on digital sovereignty and sustainability in Europe. With the world becoming increasingly connected, interdisciplinary communication is sure to play an important role in strengthening and safeguarding these values – but, as Jerker cautions, this will have its biggest impact when viewed as a tool rather than an objective. “When managers call for it, it doesn’t help because interdisciplinarity requires people to truly understand and appreciate each other. That’s the distance I spoke about previously. If the distance is too great, we can throw as many people into a room as we like and nothing will happen. But if we first find situations we might want to discuss, if we’re both lucky and skilled in forming teams and explaining our expertise and if people are open, forward-looking and innovation-minded, interdisciplinarity will work.”
“In ecosystems, you have people who are absolutely ready to play as a team and be open. You have to cherish them! But you also have experts who are very good at their technical work but are not great at thinking outside of it. These people do have a role to play in the technical details,” Philippe stresses. “For me, there are still some open questions. Today, you have the open-source community and commercially available tools, but most are not completely interoperable. There is no magic tool doing everything from A to Z, so engineers spend a lot of time creating converters. Interoperability will be absolutely mandatory if we want to push our CO2 commitment to its maximum. If we are not interoperable at every stage, we will not know our real footprint, so interdisciplinarity should be a trigger to make us more interoperable than we currently are.”
“When working in projects or creating proposals, I often still have the feeling that there’s some kind of silo thinking in the cooperation between different levels of the value chain, between disciplines and even between countries,” agrees Michael. “But things are better than in the past and people are opening up to bring together different fields. We need to go on in this way and bring the ideas of others into our own domains, even if they might seem very unconventional at first glance. We can then unlock a great potential that is currently hidden between the different parts of the ECS value chain in Europe.”