As a network of OEMs, SMEs, universities and research institutes, ARTEMIS represents over 200 organisations in ECSEL JU. Their collaborations cover a vast range of topics and domains which touch every corner of our lives – often with little awareness or recognition from society at large. In this newsletter, we’re pleased to highlight three of these many innovative projects: InSecTT, SECREDAS and BRAINE.
Can artificial intelligence (AI) support green technologies for ships? Could it be trusted for clinical decision-making in hospitals? And should we have faith in a connected vehicle? These are just some of the questions tackled by InSecTT, which brings together 52 partners in 12 countries and a budget of over 40 million euros. As a follow-up to DEWI (2014-2017) and SCOTT (2017-2020), InSecTT extends their successes in wireless connectivity and interoperability with a focus on making AI and machine learning explainable, understandable and trustable.
“The InSecTT partners believe that Artificial Intelligence of Things (AIoT) is the natural evolution for both AI and the Internet of Things (IoT) because they are mutually beneficial,” explains Michael Karner, Lead Researcher in Embedded Systems at Virtual Vehicle (Austria) and the project’s coordinator. “AI increases the value of IoT through machine learning by transforming the data into useful information knowledge, while IoT increases the value of AI through connectivity and data exchange.”
InSecTT arrives at a pertinent time. Due to the enormous complexity of IoT development, human errors are increasingly exploited by opportunistic hackers. This erodes consumer trust, which the project seeks to rebuild by transferring complex problems from the programmer to their program. In doing so, InSecTT utilises AI for two core tasks: AI-supported embedded processing for industrial tasks and AI-enhanced wireless transmission.
“We have a broad variety of domains and use-cases representing key areas of Europe’s industry,” says Michael. “Each of them has specific potential to provide intelligent, secure and trustworthy systems to make European solutions a frontrunner in cutting-edge technology.” Examples range from intelligent transportation for smart cities to cybersecurity in manufacturing, dealing not just with things that can be connected but also with moving trustable intelligence to the edge.
When the project draws to a close in 2023, success will be measured by the development and demonstration of trustable, intelligent things which are securely connected and more than just a black box that cannot be understood by users. Michael: “Another (more informal) success criteria for me is keeping the good spirit of cooperation that has been established in the predecessor projects.”
As one project gets rolling, another comes to an end. Now in its final year, SECREDAS has utilised a budget of 51 million euros and a consortium of 70 research and industrial partners to work on a reference architecture for secure and safe automated systems which is compliant with the GDPR. The need is clear: a quarter of potential buyers/users of automated driving in Europe are reluctant to take the plunge, largely due to a lack of trust in security. As NXP’s Director of Strategic Partnerships and SECREDAS project coordinator Patrick Pype explains, this need not be the case.
“SECREDAS is a ground-breaking project because it looks towards autonomous driving from security, safety and privacy perspectives simultaneously rather than from a purely technological feasibility perspective. Most companies focus only on one element, which often creates risks in the others. In this sense, SECREDAS is setting new standards on future requirements for connected and autonomous driving and also more generally on automated systems with human interaction.”
By connecting technical capabilities, different architectural approaches and individual component developments through a set of common technology elements and design patterns based on a common safety and privacy framework, SECREDAS has been able to create more than 40 integrated demonstrators. These range from sensor fusion to biometric driver authentication and will be combined into four ‘on-road’ demonstrations that aim to solve ten of the likeliest threat scenarios for future users.
For Patrick, a great source of pride has been NXP’s success in bringing together a true multi-disciplinary consortium with security and privacy experts and knowledge in the fields of automotive, railways and health. “It took a while for everyone to understand each other’s language,” he admits, “but since then it’s made tremendous progress. For the most part, we are still at the technology development stage (TRL 5-6) rather than the product development stage, but we are moving fast towards promising final results that will soon find their way into real practice. In the end, SECREDAS is all about increasing trust for future car buyers such that they feel totally safe inside connected or autonomous vehicles.”
Running from May 2020 to April 2023, the final project recognises the enormous benefits that edge computing can offer by supporting AI natively instead of in the cloud. This is more than just a software process: hardware must be designed with a specific focus on big data processing and AI. By boosting the edge framework with energy-efficient hardware and AI-empowered software systems that can process big data at the edge, BRAINE aims to be a significant enabler of long-awaited technologies across various fields.
“The BRAINE project targets use-cases for AI at the edge in healthcare, smart cities and Industry 4.0,” notes project coordinator Filippo Cugini, Head of Research Area at CNIT. “All use-cases are attractive and have concrete potential, even if they present different conditions and levels of maturity. For example, Industry 4.0 is the most mature for advance deployment, with measurable benefits in the short term. Smart cities might have great potential but regulations, costs and sometimes attitude might pose obstacles to the rapid deployment of innovative solutions. Healthcare is exciting; there is huge potential and, in the long term, it could really become a game changer for EU citizens.”
By developing new edge computing architectures and subsystems, improving data processing at the network edge and supporting scale-up and scale-out approaches, BRAINE predicts an unprecedented impact in performance and energy efficiency. This includes energy savings of 50% for control and computing functions, an 80% reduction in space and maintenance and a near 100% fault tolerance at level 5 autonomy for automated driving, robotics and mission-critical systems. The project has also pioneered a unique means of generating interaction and dissemination via AskBRAINE, a panel series in which experts discuss a number of relevant themes.
“The AskBRAINE series was proposed by our media manager Silvia Fichera and enthusiastically accepted by all project partners,” says Filippo. “The first edition [Telemetry and monitoring at the Edge] was really appreciated by both panellists and attendees. We hope to gather additional interest and attendees to make the event even more effective. I take the opportunity to invite you all to attend our next AskBRAINE event on AI at the edge, scheduled for March 25 at 3pm.”