An improved DOA estimation method for direct sequence spread spectrum communication

Direction of Arrival (DOA) estimation is a fundamental problem in array signal processing, with critical applications in radar, 5G communications, and unmanned aerial vehicle navigation. Traditional DOA algorithms, such as beamforming and subspace-based methods (e.g., MUSIC), face significant challenges when applied to spread spectrum signals, particularly Direct Sequence Spread Spectrum (DSSS), due to their wide bandwidth, low power spectral density, and pseudorandom modulation. These characteristics lead to dispersed signal energy and ambiguous phase information, severely degrading estimation accuracy. To address these limitations, this paper proposes a novel DOA estimation framework based on Canonical Polyadic (CP) tensor decomposition. By modeling the received array data as a third-order tensor (incorporating spatial, temporal, and code dimensions), CP decomposition effectively preserves the multilinear structure of the signal subspace, mitigating information loss inherent in conventional matrix-based approaches. The proposed method leverages the inherent robustness of DSSS signals against noise and multipath interference, enabling high-precision DOA estimation even in low signal-to-noise ratio (SNR) scenarios. Simulations demonstrate that the proposed method outperforms the classical MUSIC under spread spectrum communication scenario. The Root Mean Square Error (RMSE) performance improvement can be up to 3dB at low SNR.