Current heart failure(HF) examination methods suffer from complex operation and high cost

hindering early identification and home monitoring.To address this

an HF monitoring system based on multi-source sensing data and multi-domain feature fusion is proposed to achieve non-invasive

low-cost daily intelligent monitoring.First

ballistocardiogram(BCG) and phonocardiogram(PCG) are synchronously collected from two piezoelectric sensors.Then

handcrafted features are extracted

and an optimal feature set is constructed through feature screening.A residual network(ResNet) and a bidirectional long short-term memory(BiLSTM) network are connected in series to extract deep features.Finally

multi-domain features are fused

and a discriminative model is established using a multi-layer perceptron(MLP).The verification results based on the self-collected dataset demonstrate that the accuracy of this method reaches 92.18%

indicating strong potential for application transformation.This approach provides an effective technical pathway for home monitoring of HF.