Publication

Abstract : Compressed sensing (CS) is a novel idea wherein a signal can be sampled at sub-Nyquist rates and still be effectively reconstructed. Many natural signals such as the ECG signal are sparse and have sparse representation when expressed in a suitable basis. Compressed sensing exploits the sparsity by acquiring a small number of projections on to random vectors which are sufficient to recover the signal. This theory enables an effective implementation of patient-centric telecardiology or mobile cardiology systems. It guarantees improvement in performance by drastically reducing the number of samples that requires to be either stored or transmitted. Specifically in the context of ECG monitoring systems, CS signal acquisition revolutionizes miniaturization of the hardware and improves its energy efficiency. This work proposes to create a Compressed sensing based mathematical framework for the acquisition and recovery of an ECG signal which has been further classified into normal and abnormal categories. During the compression stage, a sensing matrix which results in a low PRD (Percentage root mean square distortion) is used and an efficient reconstruction algorithm is employed to retrieve the most valuable information from the ECG signal. The classification of the ECG signal is done by studying the region of interest i.e. the QRS complex. The heart rate , the interval between two R peaks and the amplitude of the R peaks contributes chiefly to determining an abnormality in the ECG signal. This model can be implemented with a wireless body sensor network (WBSN) which may be used to alert a doctor in emergency situations.