论文部分内容阅读
针对所有的基于粒子滤波的动态数据校正中粒子滤波方法仅仅是针对数据的,本文提出了一种基于粒子滤波和过程模型的动态数据校正方法(DDRPFPM)。该方法将过程模型引入到动态数据校正中,作为约束条件来更新粒子的权值,有效提高了粒子的信任度,解决了基于粒子滤波的动态数据校正(DDRPF)过程中存在的粒子退化问题。通过CSTR仿真证明,DDRPFPM能够较好的应用于动态数据校正,与DDRPF相比无论是对温度噪声还是离群值都有更强的校正能力。
For all the particle filter based on particle filter dynamic data correction method is only for data, this paper presents a dynamic data correction method based on particle filter and process model (DDRPFPM). This method introduces the process model into the dynamic data, and updates the weights of the particles as constraints to effectively increase the confidence of the particles and solve the problem of particle degeneration in the process of dynamic data correction (DDRPF) based on particle filter. The CSTR simulation proves that DDRPFPM can be applied to dynamic data correction better than DDRPF, and it has stronger correction ability both for temperature noise and outlier.