The autoregressive method is one of the simple methods for modal identification of the output-only systems. However, through the autoregressive method, for identifying n modes at least n sensors are required which is considered as the main challenging issue of this method. Also, in order to extract acceptable modal parameters using this method it is required to implement stationary response.

To overcome these issues, in this study, the empirical mode decomposition is adopted to improve the performance of an autoregressive method for extraction of frequencies and damping ratios of a structure from stationary or non-stationary response. To evaluate the efficiency of the new developed method, two building structures and support tower for a segmental bridge are considered and the frequencies and damping ratios are obtained using developed method.

The results indicated that the proposed method adequately estimates the considered parameters using stationary and non-stationary response which recorded by only one sensor. Moreover, it is proved that this method outperforms in comparison with the other relevant methods which applicable with the non-stationary responses.

自回归方法是用于仅输出系统的模式识别的简单方法之一。然而，通过自回归方法，为了识别n个模式，至少需要n个传感器，这被认为是该方法的主要难题。同样，为了使用这种方法提取可接受的模态参数，需要实现平稳响应。

为了克服这些问题，在本研究中，采用经验模式分解来提高自回归方法的性能，该自回归方法用于从固定或非固定响应中提取结构的频率和阻尼比。为了评估新开发方法的效率，考虑了两个建筑结构和分段桥的支撑塔，并使用开发方法获得了频率和阻尼比。

结果表明，所提出的方法利用仅由一个传感器记录的平稳和非平稳响应来充分估计所考虑的参数。而且，证明了该方法与适用于非平稳响应的其他相关方法相比具有更好的性能。