skip to content

Centre for Digital Built Britain completed its five-year mission and closed its doors at the end of September 2022

This website remains as a legacy of the achievements of our five-year foundational journey towards a digital built Britain
 

Osmose is a project aimed at building a new operating system for situated environments that can handle the demands of thousands of sensors and actuators that need to run in a coordinated and highly reliable fashion. The mini project facilitated the purchase of networking equipment to begin the deployment of the first prototype installation in the Computer Lab. We are also planning a prototype deployment within Pembroke College’s new Mill Lane site later in the year using the new hardware. Our software design is steadily advancing, and was published in its first prototype form in IEEE HotPost 2018.

[Final Report]

Digital infrastructure in modern urban environments is currently very Internet-centric, and involves transmitting data to physically remote environments. The cost for this is data insecurity, high response latency for interactive services and unpredictable reliability of services. This is not a sustainable future for driving the sensors that will permeate millions of buildings worldwide to a high degree of correctness.

We have been designing a new operating system -- dubbed Osmose -- that inverts the current Internet-centric architecture and provides ground-up services for connected devices in a building. Osmose inverts the current model by building an operating system designed to securely connect physical spaces with extremely low latency, high bandwidth local-area computation capabilities and service discovery.

2018 Osmose image

Applications in Osmose are designed to run simultaneously across the hundreds of embedded devices that form a modern situated environment, such as environmental sensors, audio and video capture, and also a new generation of directed sensorial generators (such as parametric 3D audio to beamform sound directly into a single individuals ears). At the same time, they require no external connectivity  to run, with local processing units providing the resources required for machine learning and storage.

We have built many of the components that comprise Osmose over the past decade, and are now embarking in an effort to deploy a real prototype in a building. Our proposed next feasibility study is to assemble a room-level instantiation of the Osmose architecture, and to test our hypothesis that room-level connectivity and user detection can be reliably performed by tiny power-efficient embedded controllers in the walls, and to run a realistic application such as user presence detection and identification via local machine learning.

[Final Report]

Researchers:


Computer Laboratory, University of Cambridge

Computer Laboratory, University of Cambridge

Computer Laboratory, University of Cambridge