A. Saad et al. (Eds.): Soft Computing in Industrial Applications, ASC 39, pp. 189–199, 2007. springerlink.com © Springer-Verlag Berlin Heidelberg 2007 AFC-ECG: An Adaptive Fuzzy ECG Classifier Wai Kei Lei 1 , Bing Nan Li 1 , Ming Chui Dong 1,2 , and Mang I Vai 2 1 Institute of System and Computer Engineering, Taipa 1356, Macau 2 Dept. Electrical & Electronic Engineering, FST, University of Macau, Taipa, Macau bingoon@ieee.org, {ma46530, dmc, fstmiv}@umac.mo Abstract. After long-term exploration, it has been well established for the mechanisms of elec- trocardiogram (ECG) in health monitoring of cardiovascular system. Within the frame of an intelligent home healthcare system, our research group is devoted to researching/developing various mobile health monitoring systems, including the smart ECG interpreter. Hence, in this paper, we introduce an adaptive fuzzy ECG classifier with orientation to smart ECG interpret- ers. It can parameterize the incoming ECG signals and then classify them into four major types for health reference: Normal (N), Premature Atria Contraction (PAC), Right Bundle Block Beat (RBBB), and Left Bundle Block Beat (LBBB). Keywords: ECG classifier; fuzzy sets; medical advisory system; health prognosis. 1 Introduction Most people do not care about their health condition until they fall in illness. Then, it is an excruciating and cost-expensive procedure for the subsequent therapy. More- over, early prevention and healthcare has been proven as an effective measure to pre- vent the sudden death due to heart diseases. So, the mode of contemporary healthcare is experiencing an ultimate revolution from disease recovery to health prevention. Nowadays home healthcare, including home health monitoring, has been widely ac- cepted to improve the quality of our life. The major advantage of home health moni- toring is to provide a cost-effective way for health prognosis with various physiologi- cal signals collected remotely [1]. In the pilot project - “Intelligent e-Home Healthcare System”, we propose a series of embedded medical advisory systems to enhance the intelligence of conventional biomedical transducers, such as intelligent sphygmogram analyzers (SGA) and smart electrocardiogram interpreters (ECGI) [2], [3], [4]. Then, people can master their health condition better at home. Beyond collecting and submitting physiological sig- nals, these intelligent healthcare apparatus, benefited from the embedded medical ad- visory system, can report the health condition in a real-time manner. Meanwhile, the embedded-link mode of medical advisory systems enable home subjects to submit the collected signals to health centers for further analysis. Coming to cardiovascular system monitoring, electrocardiogram (ECG) is most competent because it can reflect the complete cycle of subtle heart beating. An ECG signal is the record of changing bioelectric potentials with respect to time as the