Application of Symplectic Geometry Mode Decomposition Based on Gaussian Process Space Angle in DOA Estimation

In the presence of noise interference, direction of arrival (DOA) estimation becomes challenging. Particularly, when radar signals are entirely submerged in noise, the spatial information within the signals becomes obscured, thereby affecting the precision of DOA estimation. To address the issue, a symplectic geometry mode decomposition (SGMD) method based on Gaussian process space angle (GPSA) has been introduced to mitigate the impact of noise on target DOA estimation. This method operates within nonconstant curvature spaces, constructing subspaces for signals with different signal-to-noise ratios (SNRs). The subspace with the highest SNR is selected as the reference subspace. By quantifying the differences between the various signal subspaces and the reference subspace, a trajectory matrix that aligns with the dynamic characteristics of the signals is formed. Subsequently, the trajectory matrix is used to reconstruct the signals in the symplectic space, preserving the spatial information while reducing the disturbance caused by noise components on DOA estimation. Simulation and field experiments demonstrate the effectiveness of this approach in diminishing noise components within received signals, enhancing the precision of Music and Capon algorithms for DOA estimation under low SNR conditions.