Abstract:Future GW observatories, such as the Einstein Telescope (ET), are expected to detect gravitational wave signals, some of which are likely to overlap with each other. This overlap may lead to misidentification as a single GW event, potentially biasing the estimated parameters of mixture GWs. In this paper, we adapt the concept of speech separation to address this issue by applying it to signal separation of overlapping GWs. We show that deep learning models can effectively separate overlapping GW signals. The proposed method may aid in eliminating biases in parameter estimation for such signals.

What problem does this paper attempt to address?

### What problem does this paper attempt to solve? This paper aims to solve the problem of signal identification and parameter estimation bias caused by the overlap of gravitational wave (GW) signals in future third - generation (3G) gravitational - wave observatories (such as the Einstein Telescope, ET). Specifically, with the significant improvement in the sensitivity of 3G detectors, it is expected to detect more than 10^5 gravitational - wave events per year. This high detection rate increases the possibility of multiple gravitational - wave signals overlapping in time, resulting in the following problems: 1. **Signal misidentification**: Overlapping gravitational - wave signals may be misidentified as a single event. 2. **Parameter estimation bias**: When two or more gravitational - wave signals exist simultaneously, traditional parameter estimation methods may produce systematic biases, affecting the accurate estimation of the physical parameters (such as mass, spin, etc.) of these signals. ### Solution To address the above challenges, the authors propose a gravitational - wave signal separation method based on deep learning. This method draws on speech separation techniques and trains a deep neural network model to separate overlapping gravitational - wave signals. The main contributions include: - **Adaptive transfer**: Applying the concepts and techniques of speech separation to gravitational - wave signal separation. - **Deep learning framework**: Using TasNet and its variants (such as Conv - TasNet, DPRNN, DPTNet) to achieve efficient signal separation. - **Performance verification**: Verifying the effectiveness of the proposed method through simulated data sets and demonstrating its separation performance under different signal - to - noise ratios (SNR) and peak time differences. ### Formula representation The key formulas involved in the paper are as follows: - Combination of gravitational - wave signal and noise: \[ d(t)=n(t)+h_A(t)+h_B(t) \] where \(d(t)\) is the observed strain signal, \(n(t)\) is the noise, and \(h_A(t)\) and \(h_B(t)\) are two overlapping gravitational - wave signals. - Mask operation in the signal separation process: \[ \tilde{h}_A = \text{Decoder}(M_A\odot F) \] \[ \tilde{h}_B=\text{Decoder}(M_B\odot F) \] where \(F\) is the encoded feature, \(M_A\) and \(M_B\) are masks calculated for signals \(h_A(t)\) and \(h_B(t)\) respectively, and \(\odot\) represents the Hadamard product. - Overlap index for evaluating separation performance: \[ \text{overlap}(h,\tilde{h})=\frac{\int h(t)\tilde{h}(t)dt}{\sqrt{\int h^2(t)dt\int\tilde{h}^2(t)dt}} \] Through this method, researchers hope to be able to more accurately identify and analyze overlapping gravitational - wave signals in future gravitational - wave observations, thereby improving the accuracy and reliability of parameter estimation.

Gravitational Wave Mixture Separation for Future Gravitational Wave Observatories Utilizing Deep Learning

Ensemble of Deep Convolutional Neural Networks for real-time gravitational wave signal recognition

Using deep learning to predict matched signal-to-noise ratio of gravitational waves

Extraction of binary neutron star gravitational wave waveforms from Einstein Telescope using deep learning

Blind source separation in 3rd generation gravitational-wave detectors

Gravitational Wave Signal Extraction Against Non-Stationary Instrumental Noises with Deep Neural Network

Search for exotic gravitational wave signals beyond general relativity using deep learning

Deep Learning Ensemble for Real-time Gravitational Wave Detection of Spinning Binary Black Hole Mergers

Exploring gravitational-wave detection and parameter inference using deep learning methods

Parameter inference for coalescing massive black hole binaries using deep learning

Deep Residual Networks for Gravitational Wave Detection

Convolutional Neural Networks for signal detection in real LIGO data

Applying Deep Neural Networks to the Detection and Space Parameter Estimation of Compact Binary Coalescence with a Network of Gravitational Wave Detectors

WaveFormer: transformer-based denoising method for gravitational-wave data

Extracting overlapping gravitational-wave signals of galactic compact binaries: a mini review

Deep learning to detect gravitational waves from binary close encounters: Fast parameter estimation using normalizing flows

Space-based gravitational wave signal detection and extraction with deep neural network

Detection and Parameter Estimation of Gravitational Waves from Binary Neutron-Star Mergers in Real LIGO Data using Deep Learning

A Novel Deep Learning Approach to Detecting Binary Black Hole Mergers

Gravitational wave search by time-scale-recursive denoising and matched filtering

Rapid identification of time-frequency domain gravitational wave signals from binary black holes using deep learning