The Connected Things (CT) group is a development-focused group with expertise in electronics, telecommunications, and software engineering to design and develop hardware, firmware and software solutions for the IoT. CT is composed of cross-functional teams that can design, develop and deploy full-fledged IoT systems that are affordable, reliable, scalable, and intelligibly connect things to fit people's needs. CT closely works with the Human-Centred Design and Intelligent Systems groups and follows a full-stack approach. We can design and implement the different IoT layers, ranging from electronics that collect and (pre)process relevant data and connect through wireless technologies to scalable mobile and cloud services that address various application scenarios (health and industry) and follow the best practices for regulatory compliance.
How to build sustainable IoT-based solutions? By focusing on energy-optimised operation and seamless data exchange through enhanced RF communications, we can build hardware that reduces maintenance costs by exploiting energy harvesting concepts and low-powered radio communications and may even operate without battery. Given the advent of ML/AI, our hardware is designed with support for inference and distributed computing features in mind. From the software viewpoint, our approach relies on guaranteeing interoperability and dealing with heterogeneity to manage IoT devices smartly. Through standards and frameworks, and different software deployment processes (e.g., containerisation), we develop modular software architectures that enable heterogeneous devices, syntactically and semantically, to understand each other, thus overcoming vendor lock-in and allowing innovative, smart and distributed IoT applications to be built.
Edge & Cloud Computing
How to distribute computation while providing quality data and guaranteeing privacy? We rely on orchestration processes (e.g., Kubernetes) and standard topology specification languages (e.g., TOSCA) to use the IoT infrastructure (i.e., cloud, edge, and IoT devices) at its full capacity, allowing dynamic computation distribution, easing the deployment of ML/AI algorithms across different platforms and devices, and protecting users’ privacy since data is kept closer to the user. By integrating known standards (e.g., FHIR, FiWARE) or adapting formats to assure data quality, we support the development and deployment of ML/AI algorithms and data visualisation tools that help to build such algorithms used for decision-making in different application settings (e.g., healthcare, industry, agriculture). Through Blockchain, our work not only helps in the certification process of ML/AI algorithms (as data collection and sharing are traced, thus guaranteeing user privacy) but also helps improve supply chain management.
Quality Assurance & Regulatory Affairs
How to ensure sustainable technology development while guaranteeing quality assurance and regulatory pre-compliance? We oversee the entire software cycle working closely with HCD and IS teams from the requirements identification to the testing phase with the end-user and following regulatory and internal quality standards and guidelines to safeguard the development of safe, reliable, and effective technology (e.g., medical devices, ML/AI algorithms), facilitating its smooth path to market. We implement CI/CD pipelines that streamline testing, validation, and iterative improvements of ML/AI models and Software as a Medical Device (SaMD). We develop ML/AI-powered testwares that accelerate the testing process and help maintain the quality and reliability of our software products. By closely following the regulation and standard development process worldwide, we help carrying on ethical AI development practices, particularly regarding high-risk AI applications (e.g., medical devices).