Compound Interference Recognition for GNSS via Integrating Time–Frequency With Power Spectrum Features

Global navigation satellite system (GNSS) has a broad range of applications in both civil and military fields. However, the ubiquitous interference signals significantly disrupt the proper functioning of navigation equipment. Existing works have primarily focused on detecting the presence of interference in GNSS receivers or identifying a specific interference type, while the recognition of compound interference signals remains unsolved. In this paper, we propose a compound interference recognition approach for GNSS receivers via integrating time-frequency and power spectrum features. Specifically, we employ a time-frequency feature extraction module that integrates a spatial attention mechanism with convolutional layers. This module enhances the network's attention capabilities towards the salient features within the timefrequency image of compound signals. Besides, to address the limitations of time-frequency information in recognizing compound signals, we combine time-frequency and power spectrum features for interference recognition. By leveraging the correlation between different domains, we can identify multiple interference types effectively. On that basis, we develop a feature fusion network with an attention mechanism to fully capture the time-frequency and power spectrum features. Experimental results on simulated and measured datasets demonstrate that our proposed method achieves better recognition performance than baselines in various evaluation metrics.