Heart Failure Prediction Based on Continuous ECG Monitoring
HFP aims at exploring novel strategies for embedded real-time prediction and anomaly detection of major cardiac episodes, such as arrhythmias. State-of-the-art solutions in this field assume an off-line scenario, in which periodic electrocardiogram (ECG) recordings are evaluated either with automatic methods using suitable analysis software or by expert cardiologists. Instead, in this project we target implementations based on Wireless Body Sensor Nodes (WBSNs). These wearable and battery-operated devices allow for continuous and long-term monitoring of ECG signals, thus enabling a continuous, on-line assessment of the heart rhythm. On the other side of the coin, their tight performance and energy budget requires careful optimisation of algorithms and applications being executed on these platforms. While recent development in low-energy applications for features extraction of ECG characteristics has shown promising results, a complete framework for prediction of future abnormalities had not been proposed until now. The envisioned research will therefore allow a significant advancement in the treatment of cardiovascular disorders, which currently already represent the most common cause of death world-wide. The project is carried out in collaboration with the embedded systems lab (ESL) at EPF Lausanne, combining their respective expertise in embedded bio-signal analysis (ESL-EPFL) and on-line failure prediction models (USI-ALaRI), and targets an early exploration of the envisioned systems.