Key elements
Creating trust in and increasing social acceptance of intelligent systems is one of the major challenges to bring out the full potential of the IoT. To tackle this important issue, InSecTT will put a specific focus on intelligent, secure and trustworthy systems for industrial applications to provide comprehensive cost-efficient solutions of intelligent, end-to-end secure, trustworthy connectivity and interoperability to bring the Internet of Things and Artificial Intelligence together. InSecTT will not just deal with things that can be connected, but with moving trustable intelligence to the edge.
InSecTT aims at creating trust in AI-based intelligent systems and solutions as a major part of the IoT. Therefore, InSecTT will not just deal with ‘things that are connected’, but ‘trustable intelligent things that are securely and reliably connected’, i.e. moving AI to the edge and making AI and ML based systems trustable, explainable and not just a black box that cannot be understood.
Building on a sound basis – The SCOTT Trust Framework
In daily life we make many decisions that require trust, most of them we accomplish successfully. If I do not yet know the person who will take care of my daughter while I am away, what do I need to know about the person to be comfortable to entrust him or her with this task? Is the daughter of an elderly parent OK to have him living alone at his home despite his increased risk of falling? Similarly, trust plays a role in acquiring and using technical systems in economic contexts: Does the new wireless railroad-crossing system meet and exceed those of traditional wired systems? Could my personally identifiable data get into the wrong hands when I enter them on this website?
Whether a buyer or user will trust a new system depends on many circumstances that are often unknown to those who develop the system. Therefore, it becomes difficult for the designer of the individual technologies to decide what characteristics the components need to exhibit to achieve overall trust. As result of this lack of knowledge, overall trust concerns are generally delegated to the end of the system design process but not actively considered during the process.
Our approach here is that an upfront understanding of the major trust issues that a potential user or buyer will experience should facilitate the development of a better product, at potentially lower costs and significantly increased marketability. The determination of trust issues early in the design process consists of contextualizing the usage and acquisition conditions early in the design process and address them appropriately. The prerequisite is to introduce the acquisition and usage process early on and analyse them to derive potential trust issues and include requirements that would allow to address them. Also, research has identified trust enhancing design features that should help people to develop and calibrate their trust with a system. We combined these steps into a trust-enhancing framework that is depicted in the figure below.
Within InSecTT, the SCOTT trust framework will be extended towards AI and ML, with the focus of making AI and ML explainable, understandable and trustable.
The innovation metrics
InSecTT innovations will be evaluated by using DEWI-Frame [1], an innovation metrics developed within the DEWI and SCOTT projects. This assessment framework uses several interconnected layers of innovations, project requirements, project objectives and high-level objectives to derive scores for the project. The framework will be adapted to the characteristics of the InSecTT project.
The step beyond
InSecTT now goes a significant step further and will
- Bring Internet of Things and Artificial Intelligence together
- Move AI to the edge, i.e. provide intelligent processing of data applications and communication characteristics locally at the edge to enable real-time and safety-critical industrial applications
- Develop industrial-grade secure and reliable solutions that can cope with cyberattacks and difficult network conditions
- Enable AI-enhanced wireless transmission
- Provide measures for trust for user acceptance, make AI/ML explainable and not just a black box that cannot be understood
- Provide re-usable solutions across industrial domains
InSecTT utilizes a clearly use-case driven approach with 16 use cases from different areas of high relevance to European society and industry; all these use cases will be designed for a cross-domain use.
Based on the unique user-driven approach, the InSecTT project will focus on:
- A representative set of Use Cases (UC) in the different domains and Technical Building Blocks (BB) jointly enabling the demonstration of business objectives in all industrial application domains (smart infrastructure, building, production, automotive, aeronautics, railway, maritime, urban public transport as well as health).
- BBs derived from UCs (as methodologies, SW or HW components to build a SW-tool, a system or a product, as services; as profiles; as tool or tool chains; as interfaces as well as processes). The BBs are the elements in the project, where most technical work is foreseen.
- Demonstrators: Every UC driven task and every BB must contribute to a demonstrator. This approach ensures to reach the targeted Technology Readiness Level of InSecTT (TRL 7-8).
Positioning of InSecTT
InSecTT is an industry-driven innovation action (IA) in the area of trustable intelligent solutions aiming at TRL 5-8. Several research and innovation actions (RIA) such as DEWI or SHIELD, and innovation actions such as SCOTT provide a sound basis to build upon. InSecTT now is consistently going a significant step further: based on the results of the predecessor projects, InSecTT aims at local use of trustable intelligent solutions in mainly wirelessly connected elements of the “Internet of Things” for industrial applications. InSecTT is an industry-driven innovation action (IA) in the area of trustable intelligent solutions aiming at TRL 5-8. Several research and innovation actions (RIA) such as DEWI or SHIELD, and innovation actions such as SCOTT provide a sound basis to build upon.
InSecTT now is consistently going a significant step further: based on the results of the predecessor projects, InSecTT aims at local use of trustable intelligent solutions in mainly wirelessly connected elements of the “Internet of Things” for industrial applications. All the solutions of InSecTT – based on 16 industrial use cases and applications in the areas of smart infrastructure, building, production, automotive, aeronautics, railway, maritime as well as health – will be developed and finally demonstrated all over Europe. A special focus is put on the cross-domain nature of the different use cases – not only including one specific domain but going beyond with a strong focus on cross-domain applications. The cross-domain aspect is not only realised by bringing in components to different domains, but also by interconnecting the domains in a truly cross-domain communication. This can be seen e.g. on airport UCs, where information from buildings, vehicles and planes needs to be exchanged. Besides this, at the end of InSecTT a broad variety of reusable Building Blocks for reliable AI & machine learning, and components for secure, safe and reliable wireless systems will be available.
Since some aspects of trustable intelligent solutions currently still have a quite low TRL (TRL4), not only industrial partners but also relevant support from research partners is required to finally come to TRL 5-8. The table belo shows the current and target TRLs for all InSecTT Use Cases including the BBs demonstrated there.
|
Wireless Platooning communications based on AI-enhanced 5G |
AI-enriched Wireless Avionics Resource Management and Secure/Safe Operation |
Wireless Security Testing Environment for smart IOT |
Intelligent wireless systems for smart port cross-domain applications |
Smart and adaptive connected solutions across health continuum |
Location awareness for improved outcomes and efficient care delivery in healthcare |
Intelligent Transportation for Smart Cities |
Intelligent Automation Services for Smart Transportation |
Current TRL |
4 |
4 |
4 |
5 |
4 |
4 |
4 |
4 |
Target TRL |
5 |
6 |
6 |
7 |
8 |
8 |
6 |
6 |
|
Cybersecurity in Manufacturing |
Robust resources management for construction large infrastructures |
Smart Airport |
Driver Monitoring and Distraction Detection using AI |
Secure Industrial Communications System |
Secure and Resilient Collaborative Manufacturing Environments |
Intelligent Safety and Security of Public Transport in urban environment |
Airport Security – Structured and Unstructured Flow of People in Airports |
Current TRL |
4 |
4 |
4 |
4 |
4 |
4 |
5 |
5 |
Target TRL |
6 |
6 |
6 |
6 |
7 |
7 |
7 |
7 |
InSecTT TRLs per Use Case
[1] Joachim Hillebrand, Michael Karner, Werner Rom: Gaining and Keeping Overview of Complex RTDI Projects with the DEWI Assessment and Monitoring Framework (DEWI-Frame), Eleventh International Conference on System of Systems Engineering (SoSE), 2016