论文部分内容阅读
提出了一类比较普遍存在的非高斯噪声——过滤白噪声的定义 ,并且讨论了这类噪声的性质。在此基础上 ,本文拓宽了对信号噪声的要求 ,对传统的 TL S- ESPRIT谱估计算法进行了一定的改进 ,推导出了在这类噪声加性干扰下信号的广义 TL S- ESPRIT谱估计方法。通过特征分解技术 ,获得并证明了一种有实际应用价值的算法。这种算法具有高分辨、运算量低、短数据和宽频带等特点。该算法对噪声的假设更加一般化 ,具有比其他方法更强的实用性。通过一个实际的例子 ,验证了该方法的良好性能 ,说明了该方法的正确性和实用性。
A class of more general non-Gaussian noise-filtering white noise is proposed, and the properties of such noise are discussed. On this basis, the paper broadens the requirement of signal noise and improves the traditional TL-ESPRIT spectrum estimation algorithm. The generalized TL-ESPRIT spectrum estimation of the signal under such noise additive interference is deduced method. Through the eigen decomposition technique, a practical algorithm is obtained and proved. This algorithm has high resolution, low computational complexity, short data and wide bandwidth. The noise assumption is more generalized and has more practicality than other methods. Through a practical example, the good performance of the method is verified, and the correctness and practicability of the method are demonstrated.