Facts

Contact person:
Paul Davidsson
Financer:
  • The Knowledge Foundation
Responsible at MaU:
Paul Davidsson
Project members at MaU:
Collaborators :
  • Axis Communications
  • Cetetherm
  • Crossbreed
  • Ericsson
  • Infonomy
  • Sigma Technology Solutions
  • Sony Nordic
  • Sony Network Communications Nordics
Time frame:
01 October 2022 - 30 November 2025
Faculty/department:
Research environment :
Research subject:

Project description

A result of the fast developments in information technology is that more objects are becoming equipped with embedded with sensors and the ability to communicate with other objects, often referred to as the Internet of Things (IoT).

Currently many consumer products are taking this step, e.g. TVs, cameras, thermostats, and health devices. The trend is similar in other areas, where e.g. energy systems and buildings can be operated remotely and automatically optimized. The real potential often is not in the objects per se, but in the data they can provide to other services. IoT systems have several characteristics which make their design a challenging task, e.g. they are often large, data-driven, distributed, dynamic, evolving, and heterogeneous.

Moreover, different IoT applications have different desired quality characteristics, e.g. concerning response time, power consumption, usability and privacy, which influence the design. Artificial Intelligence (AI) has been proposed as a key technology to enable the realization of IoT systems meeting the requirements related to such characteristics. The integration of AI into IoT is sometimes referred to as AIoT.

To make scalable AIoT systems, the idea of utilizing distributed resources at the "edge" of the network and providing the processing capabilities closer to the source of data has been proposed. Edge computing may help to address the challenges of traditional cloud solutions for hosting IoT applications by lowering latency, handling privacy issues, and reducing data communication. Moreover, distributed edge computing can be integrated with centralized processing, resulting in hybrid architectures, combining the strengths of both approaches.

AIoT entails that IoT systems will be able to analyse data and make decisions with limited or no involvement of humans. Typically, such systems can process data and make decisions faster and more accurate than those relying on humans, but human users need to be able to trust the analyses and the decisions made by the system. Thus, how human users interact with the system becomes a key concern to support trust, e.g. through transparency and explainability.

The core research question addressed in the Synergy project can be formulated as: How should intelligent and trustworthy IoT systems be designed? In particular, we will focus on the following sub-questions:

  • How should AI be used to achieve intelligent and trustworthy IoT systems?
  • How could should edge computing be used?
  • What trust- and privacy-enhancing mechanisms are needed?

By studying this in four application areas important for a sustainable society, i.e., health, energy, surveillance, and building management, an integrated framework for supporting designers of intelligent and trustworthy IoT systems will be developed. It addresses issues concerning architectural deployment, the use of AI, and trust, while taking into consideration the relevant quality characteristics e.g. scalability and usability.