10040850Serverless Edge Computing for 5G Edge DevicesClosedCollaborative R&DInnovate UKDespite efforts to improve accident frequency rate(AFR) since 1980, construction is one of the most dangerous sectors in the UK(2.4 AFR) compared to combined industry averages(1.5). Each year, workers experience 61,000 non-fatal and 39 fatal injuries, costing £16.2bn\[HSE-2021\]. Through discussions with UK market-leading construction firm Taylor Wimpey, we have identified a clear desire/need to utilise connected edge technologies to further reduce AFR and improve workplace health&safety and associated costs. Approaches include: * IoT smart devices/wearables to monitor worker health and real-time alerting of wellness issues, e.g. physiological/environmental data, temperature/heart-rate/steps/vibration/noise. * Connected sensors to monitor site variables e.g. location/exclusion zones/temperature/pressure/dust particles, and volatile organic compounds/chemicals to send data alerts to ensure workers aren't exposed to harmful situations or materials. Edge devices require high-quality internet connections as alerts/detections need to be processed in real-time to enable effective action to minimise the impact of health&safety issues. On construction sites and other industrial environments, the lack of reliable connectivity is the main barrier to adoption. Construction firms also need to ensure that personal data is protected. Existing cloud-based solutions for edge computing require complex set-up and configuration, create potential data security issues, and are not dynamic enough to facilitate real-time data processing and achieve required latency(5ms). EMS and Loughborough University are developing a novel serverless edge FaaS (Function as a Service) engine(ServerlessEdge) to optimise existing i-MO routers to enable connected edge devices to operate onsite in real-time in traditionally low connectivity environments without the need to deploy and support additional on-site IT infrastructure or centralised services. The engine runs and terminates serverless functions on edge devices, schedules tasks, and manages and orchestrates resources to minimise latency from 30-50ms to 5ms.