King's College London

Dr Yansha Deng from the Department of Engineering is part of a €6 million three-year Horizon Europe grant, which aims to revolutionise computation and connectivity for wireless connected devices, including virtual reality and autonomous vehicles. This could enable the full-scale roll-out of computing and communication technologies for healthcare and smart cities that have not previously been possible.

The project aims to transform healthcare for patients and traffic safety in our cities through the creation of a novel computing infrastructure for wireless connected devices, powered by artificial intelligence (AI).

Entitled ‘Multi-layer 360° dYnamic orchestration and interopeRable design environmenT for compute-continUum Systems (MYRTUS)’, the system will consist of a connected computing ‘continuum’, made of separate software and hardware elements, including Cloud, Fog- and Edge-level components, to improve the latency ie. speed of internet-connected devices, including extended reality and autonomous vehicles. Using AI, MYRTUS will be able to dynamically orchestrate and manage the demands of each device and ensure that these are matched to the Cloud, Fog or Edge computation that suits their needs.

Traditionally, communication networks that used large central servers connecting to individual devices have been phased out due to their inefficiency, slower speeds and cyber security risks. Cloud, Fog- and Edge-level computing have enabled network servers to be at closer proximity to the devices using them. But these solutions so far have not offered the seamless computing, privacy protection, energy-efficiency, and low-latency requirements some applications need, as more and more people rely on the use of ‘smart’ devices like internet-enabled VR devices and driverless cars.

Solving this problem is complex, as a greater number of devices with dynamically changing needs are being connected to networks. To solve this, AI has increasingly been deployed for computation and network management capacities, and computing is shifting further away from the centre.

The researchers now hope to further develop its use in mitigating the pressures of resource allocation in real-time. By creating a ‘continuum’ system that links all the above levels of computing, powered by AI, the project aims to create a system that can address complex and dynamic needs through efficient computation and communication resource allocation, always ensuring the right balance between computing power, communication latency or speed, and proximity for devices and users. This new infrastructure can then improve the latency and energy efficiency of those IoT devices, transforming technologies in key areas, such as healthcare and traffic management.

Dr Deng, a Reader in Communications Engineering, explains, “Virtual reality (VR) technology and the metaverse is revolutionising healthcare. Internet-enabled devices in surgeries empower doctors to run accurate visual models of a patient’s illness, test different medical interventions and predict recovery times in the metaverse. This can improve patient experience by allowing for consultation in VR with multiple medical practitioners, without the need for constant travel to see them.

“In order to process all these analytics and accurate 3D visual models in real-time, there needs to be enough computational power at closer proximity to medical and VR devices. Yet this has been a stumbling block for these types of interventions and diagnostics, as devices in hospitals or homes lack this.

“Our project enables an efficient connection to the computational muscle needed to power these use cases, enabling their deployment in a way that previously wasn’t possible. By building this bedrock on which technology innovations can be rolled out, we’re contributing to more efficient and patient-centred healthcare.”

The project also holds potential for traffic management in cities. By deploying devices/autonomous vehicles connected to the internet, traffic lights and road signs could see and recognise vulnerable road users and communicate with each other to avoid collisions in real-time. For example, traffic lights could sense a child playing near a road and encourage road users to slow down in the area by switching to red.

While the concept of a ‘smart city’ where traffic management can response quickly to its environment is not new, the realisation of devices/vehicles to enable this has been held back by the complexity and large number of the data needed to compute the correct solution. MYRTUS will allow these devices to offload this stress to areas of a network that have the computational power to carry out these complex tasks, effectively enabling smart cities to come to life.

The project also promises to continue to develop responsible AI-based solutions for communication networks by providing the programming environment MYRTUS was created in. It’s hoped that by doing so, future researchers can build develop and test new technologies that can work together across proprietary boundaries and be transparent, while providing more efficient solutions to challenges like file storage for big business.