Abstract: Real-time electrocardiogram(ECG) monitoring will result in large data volume， which brings about severe pressure on the transmission and storage of the system. In order to reduce the computational complexity of data acquisition and the data volume in transmission， compressed sensing was used for the ECG signal. Taking time and error of signal reconstruction as critical indicators， the performance of different reconstruction algorithms and wavelet basis are comprehensively studied. The results demonstrate that when the compression ratio is within 30%， the percentage mean-square difference of the basis tracking algorithm is less than 4%， and the reconstruction takes the shortest time. When the compression ratio is within 70%， the error of subspace tracking algorithm is less than 10%， and the reconstruction time is always the lowest. However， the optimal wavelet basis is generally related to the specific compression ratio.

Key words: compressed sensing, electrocardiogram signal, acquisition and reconstruction, wavelet basis