- 學位論文
應用FPGA實現ECG即時監測晶片
Hardware Design of ECG Real-Time Monitoring Chip Using FPGA
摘要
為了能長時間監測慢性病或心臟病患者之心電圖(electrocardiogram, ECG)訊號，並考量其在配戴生理訊號擷取裝置時的方便性，故本文擬以現場可程式化邏輯閘陣列(FPGA)平台為基礎，設計發展一個便於患者配戴之微型ECG即時監測照護晶片。為能使此晶片具有即時監測的效果，故擬整合“So and Chan” R波偵測演算法與GreyART類神經網路分類技術，於FPGA平台上進行設計開發，以期能達到即時偵測與分類ECG訊號之目的。透過晶片化之設計可使得原本生理訊號擷取裝置積體化，進一步達到縮小體積以及降低功率損耗與成本等優點。而由實際實作模擬與分析之結果可知，本文所提出之晶片在R波偵測部份的準確率高達93.88%，且在已知R點的情況下，其ECG分類部份的平均準確率亦可達88.34%，故輔助驗證此晶片在ECG訊號即時監測方面之效能。
並列摘要
In order to monitor the electrocardiogram (ECG) signal of the patients with chronic or heart diseases, a field programmable gate array (FPGA) based ECG real-time monitoring chip for healthcare application is presented, designed, and synthesized in this thesis. The “So and Chan” R wave detection algorithm and the GreyART neural network are both applied and integrated into this chip to enable the abilities of real-time ECG beat detection and classification. Besides, comparing the proposed chip with the conventional physiological signal acquisition device, the device volume can be obviously miniaturized and easy to use for the patients, and the power consumption may be further reduced for long-term monitoring application. Finally, by the simulations and implementation results in the FPGA development platform, the accuracy rate in R wave detection part of the proposed chip is 93.88%, and the average accuracy rate in ECG classification part is 88.34% at the situation that R points are known. Therefore, the performance and feasibility of proposed ECG real-time monitoring chip can be verified.