Compressed Sampling Method Of ECG Signal Based On Random Demodulator

The wearable ECG monitor is used to collect human ECG signals to monitoring the heart functions,but the sampling process guided by the Nyquist-Shannon theory has gradually become a limiting factor in the development of health monitoring.Because the Nyquist theory states that the sampling frequency must be twice as high as the highest frequency of the external analog signals,in result the high sampling rate and large data volume have brought great challenges to the battery life,data storage,transmission and processing of the device.In recent years the compressed sensing theory has combined sampling with compression and sample the compressed information by project the signal to the sparse domain,which has great application potential in low-frequency sampling and low-power applications.Aiming at the problem of large amount of sampled data and high power consumption in ECG signal monitoring of wearable devices,the random demodulator is selected to realize compressed sensing theory,and a compressed sampling method of ECG signals based on random demodulator is proposed through theoretical demonstration,simulation and sampling experiment.The main research contents and results are as follows:(1)The compressed sensing theory and the random demodulation sampling structure are studied.Considering the compression ratio and the ECG signal reconstruction accuracy,the effects of sampling phase,filter order,cut-off frequency,single sampling time of random demodulator and sparsity parameters on ECG signal reconstruction are studied based on MIT-BIH database,and the optimal results based on different reconstruction algorithms are analyzed.Finally,The feasibility of the random demodulation theory to compress and sample the ECG signal is verified.(2)A simulation platform of random demodulation structure is built.Combine the ECG signal amplitude with sampling frequency into a structure and imported into simulink as the signal source in simulation,and call the signal generation,mixing,filtering,sampling,quantization and transmission modules to make up the random demodulation structure to finish the compressed sampling of ECG signal.As the same time the influence of pseudo-random sequence frequency on sampling frequency and reconstruction accuracy is also analyzed in simulink.The simulation show that the proposed compressed sampling method with random demodulator can sampling the ECG signal at sub-Nyquist frequency,and obtain the similar sampling effect at Nyquist frequency by reconstruction algorithm.It provides a basis for building the electric circuit of random demodulation.(3)A electrical system of random demodulation is built.The platform includes sampling circuit and software.Which the sampling circuit act the function of random demodulation to compression sampling by integrates various hardware modules with FPGA as the core.The software includes verilog synchronization control logic,testbench verification program corresponding to each control level and signal reconstruction algorithm.The sampling experiments based on MIT-BIH database show that the random demodulation circuit can compressed sampling the ECG signals in 120 Hz when the compression ratio is 10,and the reconstruction accuracy of 93-95% for different groups of ECG signals.