利用GPS系统在线监测获取的大量结构响应数据，完成了大量数据处理工作，得到了结构响应的时间序列。在此基础上，完成了结构响应的谱分析，初步建立了结构响应的状态空间模型及ARMA（autoregressive moving average method）模型。计算分析结果与大桥通车试验的实测结果及有限元模型预测结果相当吻合。研究表明，基于结构响应的状态空间模型及ARMA模型是评估结构动力性能的有力工具，同时进一步验证了通过RTK GPS系统建立结构安全监测系统的可能性。

For the huge amount of raw data that were obtained during field data acquisition by the real-time kinematic global positioning system (RTK GPS) in structural health monitoring (SHM) system of the Nanpu Bridge in Shanghai, we finished lots of processing works and obtain x-coordinate , y-coordinate,z-coordinate time series of all 12 monitoring station points. The coordinate time series were treated as the multi-outputs of the dynamic system and used to build the state space model and ARMA model and spectrum estimations for the response. The analysis results were consistent with the results of test that completed when the bridge was opened in 1990 and results that forecasted by fine finite element (FE) model. The research illustrated that the state-space model and ARMA model were efficient tools for evaluation of performance of civil structures, and furtherly proved the possibility of building the prototype SHM that based on RTK GPS.