论文部分内容阅读
本文详细分析了现有输入噪声模型和输出噪声模型的抗噪声原理,提出了一种能同时抑制输入噪声和输出噪声的无偏差估计模型,它利用二次互功率谱代替通常估计中的自功率谱。使得由于噪声引起的误差可以通过多次平均完全消除。新估计模型的精度与输入噪声和输出噪声的水平无关,仅与统计次数有关。在输入、输出噪声不相关条件下,即使相干函数很低仍能通过多次平均获得具有足够精度的频响函数值。文章最后给出了不同噪声干扰下的(?)估计值、(?)估计值以及本文模型的(?)估计值,并与没有噪声干扰下的H_o 值进行比较。
This paper analyzes in detail the anti-noise principle of the existing input noise model and the output noise model, and proposes an unbiased estimation model that can simultaneously suppress both input noise and output noise. It uses the quadratic power spectrum instead of the self-power Spectrum. So that the error due to noise can be completely eliminated by multiple averaging. The accuracy of the new estimation model is independent of the level of input noise and output noise, and is only related to the number of statistics. In the input and output noise is not related to the conditions, even if the coherence function is low can still be obtained by multiple averaging frequency response function with sufficient accuracy value. At the end of this paper, the estimated values of (?), (?) Under different noise disturbances and the estimated values of the proposed model are given and compared with the values of H_o without noise.