radarODE-MTL: A Multi-Task Learning Framework with Eccentric Gradient Alignment for Robust Radar-Based ECG Reconstruction

TL;DR

This work tackles robust radar-based ECG reconstruction under noise and movement by decomposing the problem into three interrelated tasks: morphological ECG shape recovery, ECG anchor (R-peak) detection, and cardiac cycle length estimation. It advances the methodology with radarODE-MTL, a multi-task framework that uses an eccentric gradient alignment (EGA) strategy to balance task difficulties and mitigate gradient conflicts during shared-parameter optimization. Empirical results on the MMECG dataset show that EGA improves RMSE, PCC, R^2, cycle-length accuracy, and robustness to constant and abrupt noise compared to several baselines, with the best performance around a temperature parameter T = 1.0. The approach also demonstrates competitive gains on indoor scene understanding datasets, illustrating the generality of EGA for cross-domain MTL. Overall, radarODE-MTL provides accurate, noise-robust long-term ECG reconstruction from radar signals and offers improved interpretability through task decomposition and biologically informed morphologies, with potential for practical non-contact cardiac monitoring.

Abstract

Millimeter-wave radar is promising to provide robust and accurate vital sign monitoring in an unobtrusive manner. However, the radar signal might be distorted in propagation by ambient noise or random body movement, ruining the subtle cardiac activities and destroying the vital sign recovery. In particular, the recovery of electrocardiogram (ECG) signal heavily relies on the deep-learning model and is sensitive to noise. Therefore, this work creatively deconstructs the radar-based ECG recovery into three individual tasks and proposes a multi-task learning (MTL) framework, radarODE-MTL, to increase the robustness against consistent and abrupt noises. In addition, to alleviate the potential conflicts in optimizing individual tasks, a novel multi-task optimization strategy, eccentric gradient alignment (EGA), is proposed to dynamically trim the task-specific gradients based on task difficulties in orthogonal space. The proposed radarODE-MTL with EGA is evaluated on the public dataset with prominent improvements in accuracy, and the performance remains consistent under noises. The experimental results indicate that radarODE-MTL could reconstruct accurate ECG signals robustly from radar signals and imply the application prospect in real-life situations. The code is available at: http://github.com/ZYY0844/radarODE-MTL.

Figures (9)

• Figure 1: Relationships between cardiac mechanical and electrical activities, with single cardiac cycle and ECG anchors labeled.
• Figure 2: The impact of strong noise and misalignment: (a) ECG recovery distorted by RBM noise chen2022contactless; (b) Misaligned ECG recovery due to the inaccurate PPI estimation zhang2024radarODE.
• Figure 3: Overview of the radarODE-MTL framework with EGA strategy: (a) Shared backbone extracts time-frequency features from the signal spectrograms with four layers of residual block; (b) Morphological decoder only reconstructs the shape of the current ECG piece; (c) ECG anchor decoder estimates the time-index of anchors (R peaks); (d) Cycle length decoder estimates the length of the current cardiac cycle; (e) The proposed EGA strategy for optimizing shared parameter space.
• Figure 4: Illustration of EGA: (a) Original gradient space with gradient conflict and magnitude dominance; (b) The projection of the original gradient space into the orthogonal space with equal “learning rate”; (c) The implementation of eccentric gradient alignment to skew the joint gradient $\tilde{g}_{joint}$ towards the hard task by introducing the eccentric vector $v_{ecc}$.
• Figure 5: Scenario for data collection from quasi-static subject chen2022contactless.

Theorems & Definitions (1)

• Remark 1