Department of Excellence | Research
Department of Excellence Research
Embedded Systems for Digital Medicine
The widespread use of increasingly efficient embedded systems and the simultaneous development of intelligent medical sensors and devices have made Mobile Health a crucial resource for digital medicine. Data collected from wearable, ingestible, and embedded sensors in mobile devices can be gathered and processed using Data Analytics and Artificial Intelligence techniques. This enables healthcare professionals to track user habits, monitor the progression of chronic diseases, or detect the onset of critical conditions.
In this context, DIEM’s research will develop in several directions. First, innovative algorithms will be developed for real-time analysis and interpretation of data from heterogeneous wearable sensors. This will be complemented by the study and development of architectures that fully leverage the potential of edge nodes (edge computing) to handle not only functional tasks but also data processing, analysis, correlation, and inference. It is clear that these two lines of research are highly complementary and integrated but must be accompanied by the study and development of methodologies to enhance the hardware and software security of such systems.
The integration of skills and expertise ensured by DIEM’s scientific areas will allow the development of innovative methodologies aimed at implementing reliable distributed healthcare applications and services. These will ensure both intelligent management and distribution of computational activities and effective resilience to security and reliability issues in the network, which would be highly detrimental in critical situations.
Artificial Intelligence for Digital Medicine
Artificial Intelligence and related cognitive technologies are creating disruptive opportunities in the healthcare and medical fields. Consider the possibilities offered by precision medicine, which uses Machine Learning (ML) techniques to analyze massive amounts of data, improving diagnostic capabilities and the predictability of therapeutic responses, thus enabling treatments tailored to individual characteristics.
In this area, DIEM’s research activities will focus on Machine Learning-based methods for medical imaging diagnostics, providing healthcare professionals with tools to improve the accuracy and sensitivity of diagnoses using advanced imaging technologies. Alongside this, AI methodologies will be developed for omics data analysis, including approaches based on single types of data (genomics, proteomics, metabolomics), multi-omics data integration, and the fusion of imaging and genomic data (radiogenomics).
In addition to approaches enabling the “personalization” of therapeutic and care activities, AI-based technologies that can enhance other aspects will also be considered. Specifically, text mining techniques will be developed for processing natural language medical diagnoses or documents. These techniques will enable the integration of multimodal data from various sources (medical records, clinical tests, diagnostic imaging). Using this data, systems will be built to support diagnosis and evaluate the effects of ongoing therapies.
Within the realm of Artificial Intelligence, significant importance is being placed on studying and developing technologies for continuous and personalized assistance tasks. In these cases, assistive systems, including robotic ones, must not only exhibit behaviors driven by collaborative control strategies but also be capable of interacting with patients, monitoring their condition, collaborating with medical/nursing staff, and adapting to unstructured environments.
In this domain, advanced control techniques integrated with artificial intelligence methods will be developed, enabling healthcare professionals to incorporate these systems into planning assistive activities. To make social interaction as simple and natural as possible between these systems, healthcare personnel, and patients, methodologies based on Machine Learning typical of Cognitive Robotics will be developed. These will allow high-level interaction between the patient and the system, such as recognizing the individual patient, their emotional state, and their gestures.