Mobile Embedded Health-Care System Working on Wireless Sensor Network

We proposed an ECG signal monitoring mobile embedded system, which is aimed to be utilized for a ubiquitous health-care system. The proposed monitoring system is using wireless transmission based on ZigBee protocol for signal transferring. The proposed system mainly consists of two mobile platforms. One platform is for signal acquisition and the other is for monitoring based on ECG signal analysis and recognition. A histogram of gradient algorithm is applied for general feature extraction. Moreover, multiple auto-associative multilayer perceptron neural networks with an hierarchical structure are applied for discriminating normal and abnormal ECG signals. Experimental results show that the proposed system is successfully monitoring ECG signals. In the experiments, we considered three different abnormal ECG signals such as BIDMC congestive heart failure, malignant ventricular ectopy and CU ventricular tachyarrhythmia as well as normal ECG signals obtained from MIT-BIH DB.