| Summary: | 碩士 === 國立海洋大學 === 航運技術研究所 === 84 ===
Adaptive antennas can suppress the interference efficiently, when the direction vector of the desired signal (steering vector) is correctly set. However, there exists sensor location uncertainty in practical circumstance, for example, the sensor locations are randomly perturbed, which will induce the random steering vector errors. A small perturbation error in the steering vector will cancel the desired signal as if it were a jammer. To alleviate the performance degradation caused by the steering vector errors, we propose an adaptive array based on the restoration of the Toeplitz structure of the correlation matrix and the property of the optimal weight vector within the signal subspace. By the above restoration process, the proposed array will be much insensitive to the random steering vector errors than the conventional adaptive array.
For a perfect steering, the optimal weight vector lies within the signal subspace of the correlation matrix. To alleviate the performance degradation caused by steering errors, the eigenspace-based beamformer constrains the weight vector being within the signal subspace. In this research, we first analyze the effect of random errors on the performance of the eigenspace-based beamformer. Then, we consider the property that the ideal correlation matrix of an uniform linear array is Toeplitz Hermitian. Based on this observation, we propose an adaptive array with weight vector computed by a reconstructive Toeplitz correlation matrix. Computer simulations indicate that the eigenspace-based beamformer performs better than the adaptive array with weights computed by a reconstructive Toeplitz correlation matrix for low SNR, and the situation is reverse for high SNR.
|
|---|
