by Margarete Hälker-Küsters (Fraunhofer AISEC) and Erwin Schoitsch (AIT)

Market intelligence firm IDC predicts that by 2025 each of us will be interacting with a smart device several thousand times each day. This technological change will have considerable impacts on our lives and lifestyles, and will be accompanied by new challenges, opportunities and risks. In this edition we focus on smart applications in several domains, tackle a few technological aspects and deal with security and quality issues.

The basis is laid by the Internet of Things - IoT (interactions and communication between humans and smart devices) and the Industrial Internet of Things - IIoT (in industrial context, machine-to-machine communication). However, “smart things everywhere” is not just IoT or IIoT, or mobile phones – it means intelligence, cognitive systems and technology, machine learning and artificial intelligence, security, big data and cloud connectivity, involving many domains of everyday life and a main driver for the digital transformation of our world.

International standardisation is also reacting to these evolving developments and challenges, covering technical as well ethical and societal issues. Technical aspects are handled in the Joint Technical Committee of ISO/IEC JTC1, Subcommittee SC41 (“Internet of things and related technologies”) and SC42 (“Artificial intelligence”). For IoT, the main topics are: architecture; interoperability (a key issue that “massively deployed systems work”); applications and use cases; coordination involving liaison or study groups on IoT trustworthiness (a key issue for public acceptance and liability); IIoT; societal and human factors; and blockchain security under IoT restricted resources. In the area of AI, key activities focus on a framework for AI systems using machine learning and other foundational standards; big data; and particularly on trustworthiness of AI. This covers many important issues for the applicability of AI components and systems in critical applications, like risk management, interoperability, bias in AI systems and AI aided decision making, robustness of neural networks or algorithms, but also governance and ethical and societal concerns overviews.

The European Commission as well as large organisations, e.g. in context of automated driving regulations and recommendations, have set up documents and guidelines on ethical considerations and trustworthiness of AI and computerised decision making. The most important examples are the EC High Level Expert Group’s (HLEG) Report “Ethics Guidelines for Trustworthy AI” [L1], the report of Informatics Europe and ACM Europe “When Computers Decide: European Recommendations on Machine-Learned Automated Decision Making”[L2] , and the IEEE initiative on “Ethical aligned design” (even planning certification activities) [L3].

These considerations on international activities should complement the articles reporting about applications in various domains and on quality, safety, security and risk management in general of such widely deployed systems of “smart things”. The articles are grouped into five chapters:

  • Smart Industrial Applications
  • Smart Cities, Buildings and Homes
  • Quality, Safety, Security and Risk Management
  • Smart Things Networks and Platforms
  • Other Applications of “Smart Things Everywhere”.

The keynote “Cognitive is the new Smart” by Prof. Claudia Eckert, Director of the Cognitive Internet Technologies CCIT cluster of excellence and Head of the Fraunhofer Institute for Applied and Integrated Security AISEC, shows us the way beyond the current conventional use of Internet, smart devices and IoT, towards a “Cognitive Internet”. This is a network of cognitive technologies for knowledge generation and sustainable value chains, while preserving trustworthiness, Intellectual Property Rights and GDPR (General Data Protection Regulation of the EC) compliance. This will allow to maintain Europe’s leading position in digital industry and business. CCIT, the Fraunhofer Cluster of Excellence »Cognitive Internet Technologies«, is an excellent example how to achieve these goals by joining forces in research and innovation.

On European level, particularly DG CONNECT, DG for Communications Networks, Content and Technology, who manages the EC “Digital Agenda”, the HORIZON Programmes, among them ECSEL, a Joint Undertaking and PPP (Public Private Partnerships between EC, national funding authorities and industrial resp. research partners) have taken up these challenges in their Work Programs, with research projects on Smart Manufacturing, Automated Driving, Smart Farming, Multi-Concern Assurance (Safety, Security, Performance and other Dependability properties), Smart Cities and Homes, and many traversal projects across domains and challenges. The pillars of these programs for “Digitalization of European Industry”, are IoT (physical meets digital), Big Data (value from knowledge) and AI and Autonomous Systems.

In the booklet “My agenda for Europe” of Ursula von der Leyen, the new President of the European Commission, a chapter is dedicated to “A Europe fit for the digital age”. It focuses on AI, IoT, 5G, and ethical and human implications of these technologies, empowering people through education and skills, and on protecting ourselves with respect to the risks of these technologies topics that are highlighted in the keynote of this issue of ERCIM News.

Highlights provided by the articles across domains, applications and challenges can be summarised as follows:

Industrial manufacturing processes place high demands on quality, efficiency and productivity to stay competitive. In order to reach this goal, machine learning (ML) and artificial intelligence (AI) methods are used for a wide range of data analysis, process control and production steering. ML usually needs a huge amount of data. In the production environment there are situations where only a small amount of high quality data is available, e.g. during the commissioning phase.  In this case ML provides methods known as one-shot or few–shot learning. Data in industrial processes are provided by a variety of sensors, one example is the application of acoustic monitoring for quality assurance. The data from acoustic sensors can e.g. define the right time for preventive maintenance.

In today’s digital world, large amount of data is collected from different sources. These data have to be analysed and visualised, and anomalies have to be detected within a short timeframe. Alarms and alerts have to be parsed in real time. Often self–learning algorithms are used to solve these kinds of problems. These new technologies will affect many areas, including manufacturing, financial services, healthcare, logistics, information security and also the operation of satellite constellations as well as spacecrafts.

In logistics one example is the exploitation of IoT technologies in supply chain traceability. From a wide range of devices, such as RFID, NFC, barcodes, Bluetooth, sensors, different types of data like velocity, temperature, humidity, location can be read. These data combined with ML algorithms can improve the whole supply chain for the benefit of all stakeholders. An efficient storage and pick up system based on wireless IoT technology for smart commissioning is another application field in the area of logistics.
Artificial intelligence technologies can also be applied to improve the safety of road traffic, to visualise noise pollution or to support farmers in their daily work.

Cameras monitor an intersection and send real-time and high-quality videos to a server where objects (pedestrians, cars, etc.) are identified and critical situations might be foreseen by machine learning technologies. This information will be sent to road users who can react accordingly. Based on a distributed acoustic monitoring system noise sources, noise levels and their contribution to the noise pollution at a specific location can be detected on a detailed level. Data from different weather stations are collected and analysed in a smart way to provide accurate weather forecasts (e.g. risk of freezing).

Knowledge–based medical diagnosis and therapy systems have been around since the 80s. Today, more advanced techniques enable further application areas, for example, remote diagnosis of allergic rhinitis is possible by analysing a short phrase uttered on a mobile phone. In an industrial working environment, cameras and sensors can provide information in a non-invasive way on user activities and states in order to detect health risks.

Last but not least are the smart applications we might have in the home in the future.  The article “Transforming Everyday Life through Ambient Intelligence” gives insights into our future home life.

Many of the aforementioned applications are based on a wide range of devices which are connected via a network and exchange information. To design, configure, manage and maintain these kinds of networks is challenging in terms of interoperability, scalability, energy consumption, real-time data and security. A few articles are related to this topic.
With these new technologies and developments, further challenges and risks for safety, security and privacy (users) and data arise.

A key topic is security, which has different facets. First, security by design is desirable. “Security by design” refers to the systematic collection and assessment of security goals, threats and countermeasures. Based on the results, a secure system is designed and implemented.

Many of the embedded systems do not have the same level of security features that is common in standard operation systems. This is because devices are tailored to the specific application.  A new approach, based on binary rewriting to retrofit already existing IoT systems, will be investigated in order to make the embedded systems more resilient against unauthorised access attempts. It is vital for smart systems to be protected against hacker attacks as effectively as possible.

The reliability of information in very different situations has to be ensured. This becomes crucial if, for example, personal data, contractual data, financial data and/or different stakeholders are involved. One answer is offered by blockchain technology together with the IoT and artificial intelligence. Another is the Arrowhead framework, which provides a chain of trust via mechanisms like certificates and secure on boarding. The International Data Space Association is conducting work in this area.

Systems used in real life (e.g. autonomous driving cars or medical diagnosis) have to guarantee safety, which means very high quality standards.

A characteristic of artificial and self-learning systems is that they may have an unsupervised unpredictable behaviour. Based on the criticality of the application, different quality assurance and test concepts must be developed in order to guarantee that the systems are reliable. Standards and certifications have to be evolved.

One approach to this is “cooperative risk management”, where all smart devices share their information and act together. Another approach is to develop a hierarchical model for qualities for smart objects. Based on existing quality models, new dimensions are added, which take into account the capabilities that arise with smart objects.

Links:
[L1] https://kwz.me/hEH
[L2] https://kwz.me/hEE
[L3] https://kwz.me/hEK

Please contact:
Margarete Hälker-Küsters
Fraunhofer-AISEC, Germany
This email address is being protected from spambots. You need JavaScript enabled to view it.

Erwin Schoitsch
AIT Austrian Institute of Technology, Austria
This email address is being protected from spambots. You need JavaScript enabled to view it.

Next issue: January 2020
Special theme:
Educational Technology
Call for the next issue
Image ERCIM News 119 epub
This issue in ePub format

Get the latest issue to your desktop
RSS Feed