The PRISM project collaborates with the NHS and UK Dementia Research to provide a simple and intuitive interface for users in the health and medical domain for monitoring and home care solutions, while providing a privacy-preserving data collection mechanism.
Internet of Things devices are increasingly being used in the health and medical domain, providing monitoring and home care solutions for a range of conditions including elderly care and monitoring. These devices often come with a range of sensors and actuators, require access to numerous personal data sources and continuous internet connectivity, and are equipped with a variety of embedded pre-trained Machine Learning models. Churning through a large variety of data sources, these models, interactions with them, and their inferences, directly and indirectly affect our lives. Recently, Covid-19 pandemic has led to long term sustained change in our lifestyle, with many of us spending more time working from home, leading to a rise in adopting IoT devices and always-on connectivity. The data from these devices is often available through the cloud-based solutions provided by the manufacturers, with varying resolution and aggregation strategies. With an increasing number of these devices being present in our homes and a larger number of information leakage channels, there is an increasing range of security threats and privacy risks.
The PRISM research team use findings from existing research and knowledge gained from relevant PETRAS project. They propose to translate their lab-based platform into a home gateway setting, ready for real-world trials. We will adopt lightweight ML models from our advanced IoT labs, integrate the crowdsourced blocking list from IoTrim, and enable privacy-preserving device behaviour analytics gathering through Federated Learning. Their approach enables translation of lab-based experiments, published in respectable international venues, into a prototype on a home gateway to be evaluated with patients taking part in large-scale trials.
The PRISM poster displayed at the PETRAS Academic Conference | Networking Research Showcase on 16 June 2022: