Robust Techniques for Designing Remote Real-Time Arrhythmias Classification System

摘要

This work describes a robust methods used in the development of "Remote Real-time Arrhythmias Classification System for multi-patients" (RACSM) to assist the medical diagnosis of cardiac arrhythmias. A coder based on a function of linear predictors and Huffman coding was chosen and found to give a compression ratio of at least 2:1 and as high as 3:1 on real-world Electrocardiogram (ECG) signals. Four techniques have been used to achieve a reduction in Hardware energy consumption of 43.30%. For setting up the communication link for Multi-Patient; a Wireless Sensor Protocol (WSP) was proposed and implemented using Time Division Duplex (TDD) scheme. A robust method to accurately classify Cardiac arrhythmias through a combination of wavelets and fuzzy logic has been introduced. The method performed well when evaluated with the MIT-BIH Arrhythmia Database. At last, the algorithms and methods mentioned above are applied in the (Two-patient) ECG Remote diagnosis system to classify and diagnose 13 kinds of arrhythmia.