Signal modeling is an important technique in many engineering applications. This paper is concerned about signal modeling problem for the sine multi-frequency signals or periodic signals. In terms of different characteristics between the signal output and the signal parameters, a separable modeling scheme is presented for estimating the signal parameters. In order to seize the real-time information of the signals to be modeled, a sliding measurement window is designed for using the observations dynamically and implementing accurate parameter estimates. Because the amplitude parameters are linear with respect to the signal output and the angular frequency parameters are nonlinear with respect to the signal output, the signal parameters are separated into a linear parameter set and a nonlinear parameter set. Based on these separable parameter sets, a nonlinear optimization problem is converted into a combination of the optimization quadric and the nonlinear optimization. Then, a separable multi-innovation Newton iterative signal modeling method is derived and implemented to estimate sine multi-frequency signals and periodic signals. The simulation results are found to be effective of modeling dynamic signals. For the reason that the proposed method is based on dynamic sliding measurement window, it can be used for online estimation applications.

信号建模是许多工程应用中的一项重要技术。本文关注的是正弦多频信号或周期信号的信号建模问题。针对信号输出与信号参数之间的不同特性，提出了一种可分离的建模方案来估计信号参数。为了捕捉待建模信号的实时信息，设计了一个滑动测量窗口，用于动态地使用观测值并实现精确的参数估计。因为幅度参数相对于信号输出是线性的，而角频率参数相对于信号输出是非线性的，所以信号参数被分为线性参数集和非线性参数集。基于这些可分离的参数集，非线性优化问题转化为优化二次方程和非线性优化的组合。然后，推导并实现了一种可分离的多创新牛顿迭代信号建模方法，用于估计正弦多频信号和周期信号。发现仿真结果对动态信号建模是有效的。由于所提出的方法基于动态滑动测量窗口，因此可用于在线估计应用。发现仿真结果对动态信号建模是有效的。由于所提出的方法基于动态滑动测量窗口，因此可用于在线估计应用。发现仿真结果对动态信号建模是有效的。由于所提出的方法基于动态滑动测量窗口，因此可用于在线估计应用。