An Improved Complex-valued FastICA Algorithm for Jamming Signals Sorting in Beidou Navigation Satellite System

Guangshun Xie,Huaiyu Tang,Rui Xue

DOI: https://doi.org/10.1109/icicsp50920.2020.9232049

2020-01-01

Abstract:The minimum level of the signal received by receiving antennas is extremely weak in the Beidou navigation satellite system (BDS) due to distance attenuation. The jamming of different modulation types directly affects the performance of the navigation receiver or even prevent its normal operation. Sorting the jamming of different modulation types at the receiving end will help the anti-jamming design of the Beidou receiver, thereby improving the receiver's anti-jamming performance. The complex-valued fast independent component analysis (FastICA) is unsuitable for sorting multiple jamming signals under the condition of low signal-to-noise ratio (SNR). Thus, this paper proposes an improved complex FastICA (c-FastICA) algorithm. First, the noise channel is introduced in the observation signals, and pseudo-whitening is performed. Then, the noise factor is introduced in the update iteration of the separation matrix to form a new iterative formula. Subsequently, the separation matrix is solved through Newton iteration. Finally, the separation matrix is multiplied by the preprocessed mixing matrix to obtain the separated signals. Theoretical analysis and simulation results show that compared with the denoising c-FastICA algorithm, the proposed algorithm greatly improves the separation effect in the case of low SNR and has a lower and more stable Amari index.