关键词: 序列二次规划；时间序列；动态时间弯曲；弯曲路径；动态可局部称重

Abstract:

To realize weighting in motion, a method for dynamic locally weighting based on sequential quadratic programming was put forward. The real-time measured signals were defined as numerical time series of
measured values and its corresponding symbolic time series of theoretical
values was solved by taking the moving force and simply supported beam as weighting physical model. The path between the two time series was extended from Euclidean distance path to dynamic time warping path. The recursive computation mathematical model of the dynamic locally weighting was established via taking the minimum of the square of deviance between the two series along the warping path as the goal. The sequential quadratic programming algorithm was employed to obtain the each moving force by means of solving the constrained nonlinear optimization model. The experimental results prove that the computation precision and calculation efficiency of the method are
favorable enough for the practical requirements.

Key words: sequential quadratic programming, time series, dynamic time warping, warping path, dynamic locally weighting