DOA Spectrum Peak Estimation Using Spatial Smoothing Method

Spatial smoothing is an effective signal processing technique for solving direction-of-arrival (DOA) estimation problems involving coherent signals. Traditional DOA estimation methods like MUSIC (Multiple Signal Classification) algorithm typically assume non-coherent signal sources. However, in practical scenarios, multipath effects or correlated interference can enhance signal coherence, thereby degrading the performance of conventional algorithms. Spatial smoothing technology improves covariance matrix structure to effectively restore rank deficiency and enhance DOA estimation accuracy.

### Fundamental Principle The core concept of spatial smoothing involves dividing a uniform linear array (ULA) into several overlapping subarrays. Each subarray's covariance matrix is computed independently, then these matrices are averaged to construct a "smoothed" covariance matrix. This process eliminates signal coherence, enabling traditional subspace-based algorithms (such as MUSIC or ESPRIT) to remain applicable.

### Implementation Steps Subarray Partitioning: Divide the original array into multiple overlapping subarrays, typically with subarray length smaller than the total number of array elements to ensure sufficient smoothing iterations. Implementation note: In MATLAB, this can be achieved using array indexing operations like subarray = full_array(start_index:end_index). Covariance Matrix Calculation: Compute the covariance matrix for each subarray's received data, preserving spatial characteristics of signals. Code approach: Use matrix operations R_sub = (X_sub * X_sub')/N where X_sub represents subarray data matrix. Smoothing Processing: Average all subarray covariance matrices to obtain a new rank-restored covariance matrix. Algorithm detail: R_smooth = mean(R_sub1, R_sub2, ..., R_subK) across K subarrays. Spectrum Peak Search: Perform spectral peak search using smoothed covariance matrix through algorithms like MUSIC to estimate signal DOAs. Key function: Implement eigenvalue decomposition and spatial spectrum calculation using functions like eig() or svd().

### Advantages and Limitations Advantages: Effectively handles coherent signals, improves DOA estimation robustness; suitable for common array structures like uniform linear arrays. Limitations: Smoothing process sacrifices some array aperture, potentially affecting resolution; computational complexity increases, particularly in large-array scenarios.

Spatial smoothing method provides crucial solutions for coherent signal DOA estimation, with wide applications in wireless communications, radar systems, and acoustic positioning fields.