Abstract:Suppression of the Power-Line Interference (PLI) in electrocardiogram (ECG) has been the key to accurately diagnosing the heart condition. A novel method to suppress PLI is presented to preprocess ECG signals. The extraction of PLI is accomplished by a trained model based on Recurrent Neural Networks (RNN), with Gated Recurrent Unit (GRU). The RNN-based model is trained to adapt the amplitude and phase of a 50-Hz sinusoidal signal to approximate the PLI in the ECG signals. The ECG signals are then filtered by subtracting the extracted PLI signals. The proposed method could significantly relieve the distortion of the ECG signals, especially in the scope of QRS complex. Detection of the existence of PLI is unnecessary because the extracted signals from PLI-free ECG signals by the trained model are negligible. Experiments are conducted on both synthetic and real ECG signals. The proposed method was evaluated by comparing with a state-of-the-art filtering method based on Kalman smoother and a traditional IIR notch filter. The results show that our proposed method improves the output SNR value by 10.5% in comparison with the referenced Kalman smoother method, and more than 20.9% and 16.3% reduction of the settling time of step increase and decrease response has been achieved, respectively. The results show that the model has enough generalization ability for unseen signals without retraining. The sequentially input and output architecture of the model makes it suitable for real-time applications and hardware implementation.

Digital Predistortion of RF Power Amplifiers With Phase-Gated Recurrent Neural Networks

A Uniform Neural Network Digital Predistortion Model of RF Power Amplifiers for Scalable Applications

Power-line Interference Suppression in Electrocardiogram Using Recurrent Neural Networks.

Wavelet Network Based Predistortion Method for Wideband RF Power Amplifiers Exhibiting Memory Effects

Convolutional Neural Network for Behavioral Modeling and Predistortion of Wideband Power Amplifiers

Augmented Convolutional Neural Network for Behavioral Modeling and Digital Predistortion of Concurrent Multiband Power Amplifiers

A Cascaded Memory Polynomial-Neural Network Behavior Model For Digital Predistortion

MP-DPD: Low-Complexity Mixed-Precision Neural Networks for Energy-Efficient Digital Predistortion of Wideband Power Amplifiers

Behavioral Modeling of Nonlinear RF Power Amplifiers Using Ensemble SDBCC Network

Walsh-domain Neural Network for Power Amplifier Behavioral Modelling and Digital Predistortion

Digital Predistortion Technique and Its Development

A Memorized Recurrent Neural Network Design for Wide Bandwidth PA Linearization

Behavioral Modeling and Digital Predistortion for Power Amplifier Based on the Sparse Smooth Twin Support Vector Regression Method

A Wideband Behavioral Model with Multiple States for RF Power Amplifier Based on Improved Recurrent Neural Network

A Novel Digital Predistortion Coefficients Prediction Technique for Dynamic PA Nonlinearities Using Artificial Neural Networks

Behavior Modeling and Digital Predistortion of Mismatched Wireless Transmitters Using Convolutional Neural Networks.

Efficient attention guided 5G power amplifier digital predistortion

Continual Learning Digital Predistortion of RF Power Amplifier for 6G AI-Empowered Wireless Communication

A Low Running Complexity Model for Digital Predistortion of RF Power Amplifiers

Behavioral Modeling of Wideband RF Power Amplifiers Using Complex-valued Wavelet Networks