基于不同约束条件的动态时间错位理论及应用

A Study on Dynamic Time Warping Based on Different Constraints

在两种不同模式的匹配过程中 ,研究了基于不同约束条件的动态时间错位理论及应用。分别采用始终点约束、不同的局部连续约束 ( Itkura约束、Sakoe- Chiba约束 )、不同的全局约束( Itkura约束、Sakoe- Chiba约束 )条件对间歇反应过程中两批次轨迹进行合理匹配 ,并使两轨迹基于动态时间错位理论获取其间的最短标准总体距离 ,在此基础上搜寻最优匹配路径 ,使两种持续时间不等的轨迹达到模式一致。应用聚合反应中的数据轨迹 ,表明此方法的实施过程和有效性。

This paper studies dynamic time warping (DTW) based on different constraints in matching two different patterns. Two trajectories in batch process can be matched reasonably in endpoint constraints, different local continuity constraints (Itkura constraint, Sakoe Chiba constraint) and different global constraints (Itkura constraint、Sakoe Chiba constraint) respectively, and minimum normalized total distance between two trajectories can be obtained based on dynamic time warping, then optimal path can be found to reconcile two different trajectories. Data trajectories from an industrial polymerization reactor are used to illustrate the implementation and the performance of this method.