Sensing & AI

Emergent Things

Networked embedded sensor-based systems can self-repair and self-update, in terms of both their hardware and software infrastructures. In doing so, this initiative investigates new ways to co-specify systems architectures, learn new software-hardware designs for 3D printing, and evolve new operating systems.

Traditional computer systems deploy hardware and provide tools to enable self-optimisation and update. But this evolution dynamic is in software – the systems does not evolve its hardware. The Emergent Things programme, leveraging connections in the departments of Computing, EE, and Design, are investigating a wholly new approach to system design that both emerges its architecture, hardware, software, and indeed the intelligence infrastructure to underpin this. Our initial focus is on Networked embedded sensor-based systems (NESS). While in-situ a basic NESS can evolve its improved operation through learning its context, etc. This clearly requires AI to learn and then enact changes, and indeed learn from the changes made. This is already difficult in software-based systems, hardware brings a real challenge. To address this a more lateral view is required to answer major research questions like: How do we translate what the user wants to the system that is (self) delivered? How do we design devices that are not only printed on demand, but have onboard printing facilities to update hardware? What does an Operating System for this even look like? Is a cross-layer communications approach the best abstraction to achieve agility, resilience, and longevity? Can this system’s intelligence grow and update, can it self-protect? Is that a good thing or a bad thing? In answering these, we will explore a completely new way to design computers.

Led by Professor Julie McCann, Dr Dalal Alrajeh and Dr David Boyle.