Speaker
Description
The data analysis of future space-based gravitational wave detectors like LISA and Taiji face significant challenges due to non-stationarities in their data, originating from time-varying astrophysical confusion foregrounds and instrumental noise drifts, which compromise traditional Fourier-domain analysis methods. In this work, we address this challenge by proposing a unified formalism based on Short-Time Fourier Transform (STFT), in order to enhance the estimation of Galactic binary (GB) signal parameters and instrumental noise characteristics admist non-stationarities. Our approach segments data into locally stationary intervals, incorporates windowing to mitigate spectral leakage, derives GPU-accelerated waveform templates and noise spectral models compatible with Taiji’s realistic orbits, and perform parameter estimations using time-frequency domain statistics. Tests on simulated data via a ”noise-agnostic workflow” demonstrate that, compared to conventinoal Fourier-domain methods, our STFT approach yields more unbiased estimation for GB signals, and provides tighter constraints on instrumetal noises. These results confirm the framework’s robustness against noise drifts (and potentially data gaps) while maintaining computational efficiency, highlighting its potential for future global analysis pipelines, especially in discriminating anisotropic stochastic GW backgrounds and instrumental noises.
