TECoSA addresses a research area where today some initial research efforts are visible, while this area will become much more important over the next decade.
TECoSA will establish basic results where the specifics of edge computing systems can be leveraged in the general context of trustworthy systems and applications. We therefore believe that TECoSA will open up several research avenues like safe machine learning, low-latency real-time compute and communications, secure and trustable compute and storage and others in this strategic research domain, and thus will trigger the academic and industrial community to pick up research results and approaches, improve them and transpose them into different system and application contexts. Most importantly, the TECoSA consortium sees a window of opportunity to initiate this research now, to maximize the research impact.
Areas of relevance
The industrial partners in the TECoSA consortium all envision edge computing applications to be of vital importance to their future business. The following areas will be of specific interest.
Safety-critical interactions between humans and cyber-physical systems
With respect to interactive human-in-the-loop applications, several industrial partners are interested in realizing new products/services like safety-critical guidance of humans in manufacturing, medical and maintenance scenarios, enabling them to commercialize much more efficient and flexible products. However, due to the safety and security constraints mentioned above, these companies face substantial challenges in realizing these novel products.
The same is true for more general safety-critical interactions between humans and cyber-physical systems, like robots or autonomous cars, through gestures or spoken dialogues. Realizing corresponding products with respect to shared workspaces between robots and humans or in public environments like street intersections, is also of interest to several partners. However, the same concerns exist with respect to safety and security leading to the same challenges for the companies.
Finally there is a large market for such applications based on edge computing infrastructures, including for consulting expertise in this respect.
Data-driven real-time analytics
With respect to data-driven real-time analytics, a major concern of industrial partners is reliability, security and trust, in particular with respect to collecting state information from the involved devices and machinery. A typical use case for this is to utilize spare processing capability to perform predictive maintenance.
However, orchestrating this distributed capability as well as securely interacting with the cloud leads to major concerns. Furthermore, the robustness of machine-learning approaches to adversarial input is a major issue. Real-time analytics is of major interest in all industrial domains covered in TECoSA. Security concerns then needs to be overcome with respect to distributed processing, and anomaly detection provided, for instance to realize intrusion detection for general cyber-physical systems.
Collaborative cyber-physical systems
Several partners wish to realize collaborative cyber-physical systems by means of edge computing, for example relating to smart manufacturing and safe automation. However, typically either guarantees of communication and computation infrastructures are missing for edge computing, or there are challenges with respect to the application of machine learning in such scenarios, effectively keeping these companies from introducing new services.
There is further a need to work out ways for certifying the corresponding safety-critical CPS, which requires fundamentally new approaches. The predictability of edge infrastructures is a concern for several of the partners with interests in providing software/hardware for safety functionality of CPS in edge computing contexts.
Finally, there are shared interests in drone fleet coordination via edge infrastructures, including for search and rescue missions. They face challenges with respect to the reliability of systems, and safety and security concerns.
Edge computing infrastructures
Beyond application-driven cases, several partners are generally interested in devising reliable, safe and secure edge computing infrastructures – or components thereof. There is a fundamental challenge in migrating from device-centric management of networks to a data-centric or application-centric paradigm. Furthermore, upcoming 5G networks are challenged by latency, reliability and security requirements to be fulfilled for all the future applications discussed above, which raise vital research and development needs.
Strategic collaboration areas
The Swedish government has further defined five strategic collaboration areas ”samverkansområden”: transportation, smart cities, circular economy, life science, and connected industry and new materials. Progress in these areas depends crucially on digitalization. TECoSA is of relevance and important for all these Swedish strategic collaboration areas. The TECoSA consortium also includes partners with interests in all these collaboration areas. Finally, and importantly, all the partners share the vision to contribute to the UN Sustainable Development Goals (SDGs). This is explicitly stated as part of the visions and goals of the large industrial partners and also for several of the SMEs.
KTH’s development plan highlights the goal to be a leader in digitalization and sustainability. TECoSA will act as an enabler for the industrial partners to realize their visions for the SDGs, in particular related to SDG goal Industry, innovation and infrastructure. TECoSA will also contribute to SDG goals Good health and well-being, Quality education, Decent work and economic growth, Sustainable cities and communities, and Responsible production and consumption.