Compressive sensing and applications

Compressive Sensing is an emerging signal processing pardigma that enables sub-Nyquist processing of sparse signals. It is based on the assumption that a small collection of linear projections (measurements) of a sparse signal contains enough information for reconstruction. The purpose of our work is to investigate compressive sensing techniques in the framework of bio-signal acquisition and classification. In particular, Electrocardiogram (ECG) signals are widely used in health monitoring and conventional ECG systems are restricted by patient's mobility, transmission capacity and physical size. Wireless Body Area Networks (WBANs) promise to be a key element in wireless ECG systems. WBANs consist of biomedical wireless sensors attached on or implanted in the body to collect vital biomedical data from the human body providing continuous health monitoring. Power consumption limitations determine that the amount of transmitted or stored data needs to be reduced without degrading the information quality. To this purpose, Compressive Sensing (CS) can be applied to bio-signal acquisition in order to compress and possibly classify the acquired data. Sparsity of the ECG signal can be exploited using a specific overcomplete dictionary, which ensures perfect reconstruction with few random measurements. The investigation developed so far has also considered a framework to directly classify compressed ECG samples into normal and abnormal states.
PhD ExpO Year: 
Research Area: 
Ingegneria delle Telecomunicazioni