Adaptive Down Sampling by Improved Methods of FRI

In modern signal processing, sampling is a key step. The number of sampling points directly affects the computation of subsequent signal processing. Nyquist sampling theorem uses twice the highest frequency of signal to sample the signal. In fact, for sparse signal, not all of the points are necessary. FRI is a highly efficient sampling method, but FRI can merely deal with discrete signals. By the improved methods of FRI, FRI theory can be extended to process continuous ECG signals. What’s more is, sampling scheme put forward by this paper can change the number of points according to the application. If the requisite degree of accuracy is low, less points are needed. Finally, simulation experiment shows that this method can not only reduce sampling rate greatly, but also can ensure the accuracy of recovery.